U.S. patent application number 15/620611 was filed with the patent office on 2018-01-25 for system, method and applications involving identification of biological circuits such as neurological characteristics.
The applicant listed for this patent is The Board of Trustees of the Leland Stanford Junior University. Invention is credited to Raag D. Airan, Karl Deisseroth, Leslie A. Meltzer.
Application Number | 20180020921 15/620611 |
Document ID | / |
Family ID | 39690530 |
Filed Date | 2018-01-25 |
United States Patent
Application |
20180020921 |
Kind Code |
A1 |
Deisseroth; Karl ; et
al. |
January 25, 2018 |
SYSTEM, METHOD AND APPLICATIONS INVOLVING IDENTIFICATION OF
BIOLOGICAL CIRCUITS SUCH AS NEUROLOGICAL CHARACTERISTICS
Abstract
Various aspects are directed to systems and methods for
assessing neural activity of a neural region having multiple
subfields. In certain embodiments, a method includes evoking a
cellular electrical response in at least one subfield due to neural
activity in the neural region, capturing image data of the
electrical response at a level sufficiently detailed in space and
time to differentiate between polarization-based events of two
respective portions of the subfield, and then assessing neural
activity by correlating space and time information, from the
captured data, for the two respective portions of the sub-field.
Other more specific aspects of the invention involve different
preparation and neural stimulation approaches which can vary
depending on the application.
Inventors: |
Deisseroth; Karl; (Stanford,
CA) ; Airan; Raag D.; (Menlo Park, CA) ;
Meltzer; Leslie A.; (Palo Alto, CA) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
The Board of Trustees of the Leland Stanford Junior
University |
Stanford |
CA |
US |
|
|
Family ID: |
39690530 |
Appl. No.: |
15/620611 |
Filed: |
June 12, 2017 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
13763132 |
Feb 8, 2013 |
9693692 |
|
|
15620611 |
|
|
|
|
12031651 |
Feb 14, 2008 |
8401609 |
|
|
13763132 |
|
|
|
|
60901178 |
Feb 14, 2007 |
|
|
|
Current U.S.
Class: |
250/307 ;
250/309; 600/407 |
Current CPC
Class: |
G01N 21/6452 20130101;
G01N 21/21 20130101; G01N 2021/218 20130101; A61B 5/4848 20130101;
A61B 5/0059 20130101; G01N 2021/217 20130101 |
International
Class: |
A61B 5/00 20060101
A61B005/00; G01N 21/21 20060101 G01N021/21 |
Claims
1.-38. (canceled)
39. A system for assessing neural activity, the system comprising:
a stimulator configured to evoke a cellular electrical response in
at least one subfield of a neural tissue in a neural region having
multiple subfields by stimulating neuronal cells of the neural
region that express a light-activated ion channel and/or a
light-activated ion pump protein; an imager configured to capture
image data of the cellular electrical response at space and time
limits of resolution that allow for differentiation between
polarization-based events of two portions of the at least one
subfield, wherein each of the two portions comprises a group of
neurons; and a processor configured for assessing neural activity,
wherein assessing neural activity comprises correlating space and
time information, from the captured image data, for the two
portions of the at least one subfield to produce correlation
results.
40. The system according to claim 39, further comprising a
composition comprising a nucleic acid encoding the light-activated
ion channel and/or the light-activated ion pump protein.
41. The system according to claim 39, further comprising a
composition for staining the neural tissue, wherein the composition
comprises at least one of: a voltage sensitive dye, and a calcium
sensitive dye.
42. The system according to claim 39, wherein the correlation
results comprise a correlation between a signal strength of a
stimulation profile and each pixel in the image data, and wherein
the processor is further configured for: determining the
correlation between the signal strength of the stimulation profile
and each pixel in the image data; determining a cross-correlation
amplitude for each pixel based on the determined correlation;
determining a cross-correlation phase for each pixel based on the
determined correlation; identifying areas of interest as a function
of a spatial location of each pixel and the determined
cross-correlations; and limiting pixel variations in the image data
that are not in the identified areas of interest.
43. The system according to claim 39, wherein the processor is
further configured for: assessing target locations for a treatment
that includes a physical intervention.
44. The system according to claim 39, further comprising a computer
arrangement comprising: the processor; and a storage medium storing
computer-executable data, wherein, when the computer-executable
data is executed by the processor, the processor assesses neural
activity by correlating space and time information, from the
captured data, for the two portions of the at least one
subfield.
45. The system according to claim 39, wherein the processor is
configured for comparing the correlation results with and without a
treatment for the neural region, wherein the treatment is at least
one of: a pharmacological chemical based substance, a
nonchemical-based therapeutic treatment, a neural-invasive
treatment, a neural-genesis treatment, and a neural-modulation
treatment.
46. The system according to claim 39, wherein the processor is
configured for assessing endophenotypes for a predictive value
relative to a disorder.
47. The system according to claim 39, wherein assessing neural
activity further comprises assessing a role of one or more
subfields within a neural circuit.
48. The system according to claim 39, wherein neural activity is
assessed relative to a depressed state of the neural region.
49. The system according to claim 39, wherein neural activity is
assessed relative to one or more of: Alzheimer's disease, mild
cognitive impairment, autism, bipolar disorder, schizophrenia, and
Down Syndrome.
50. The system according to claim 39, wherein: the space limit of
resolution is less than a centimeter; the time limit of resolution
is less than 500 milliseconds; and the polarization-based events
involve at least one of: depolarization events and
hyperpolarization events.
51. The system according to claim 39, wherein the space limit of
resolution is less than about a millimeter.
52. The system according to claim 39, wherein the subfields include
dentate gyrus (DG) and CA1 subfields of the hippocampus.
53. The system according to claim 39, wherein the neural region is
selected from the group consisting of hypothalamus, frontal cortex,
entorhinal cortex, cingulate cortex, mammillary bodies, septum, bed
nucleus of stria terminalis, amygdala, and accumbens.
54. The system according to claim 39, wherein the light-activated
ion channel and/or the light-activated ion pump protein is at least
one of: a channel-rhodopsin (ChR2), and a halorhodopsin (NpHR).
55. The system according to claim 39, wherein the neural tissue is
a brain slice.
56. The system according to claim 39, wherein the processor is
configured for correlating an activity level of the at least one
subfield from the captured image data, and providing processed
correlation results useful for assessment of the response of the
neural region.
57. The system according to claim 41, wherein the composition for
staining the neural tissue comprises the voltage sensitive dye.
58. The system according to claim 39, wherein the stimulator is
configured to evoke the cellular electrical response by optically
stimulating the neuronal cells.
59. The system according to claim 39, wherein the stimulator is
configured to evoke the cellular electrical response by
electrically stimulating the neuronal cells.
60. The system according to claim 39, wherein the stimulator is
configured to evoke the cellular electrical response by using
infrared light.
61. The system according to claim 42, wherein the processor is
configured for limiting pixel variations in the image data that are
not in the identified areas of interest by median filtering the
image data.
62. The system according to claim 61, wherein the processor is
configured for median filtering the image data by median filtering
with a 3.times.3 pixel window.
Description
RELATED PATENT DOCUMENTS
[0001] This patent document claims the benefit, under 35 U.S.C.
.sctn.119(e), of U.S. Provisional Patent Application Ser. No.
60/901,178, entitled System, Method and Applications Involving
Identification of Neurological Characteristics and filed on Feb.
14, 2007; this patent application, including the Appendix therein,
is fully incorporated herein by reference.
FIELD OF THE INVENTION
[0002] The present invention relates generally to imaging systems
and methods, and more particularly to imaging of biological
networks.
BACKGROUND
[0003] Understanding the complexities of biological tissue and its
electrical behavior continues to be an area of ongoing research and
study. For example, a major challenge facing psychiatry is the lack
of understanding of the neuronal network changes that underlie
clinical depression and recovery. The hippocampus is hypothesized
to play a central role in both depression pathophysiology and
treatment response, but the underlying local network dynamics are
not understood, with studies yielding apparently contradictory
findings.
[0004] Development of new treatments for psychiatric disorders is
hindered by an almost complete lack of information on how
maladaptive neural physiology may give rise to affective
phenotypes. For this reason, identification of a neurophysiological
final common pathway linked to the etiology of a psychiatric
disease could revolutionize understanding and guide clinical
development of novel treatments. In depression, a leading cause of
disability that affects an estimated 121 million people worldwide,
the current widely-used medications are often effective in reducing
symptoms and can promote remission, but treatment-resistance to
first-line antidepressants like the selective serotonin reuptake
inhibitors (SSRIs, such as fluoxetine and paroxetine) occurs in up
to 50% of patients. Well-known medication side effects further
complicate compliance and recovery, pointing to the need for new
classes of treatment. Development of new classes of treatment is
severely constrained by the incomplete understanding of the
multifactorial biological etiology of depression, which involves
genetic predisposition, epigenetic and developmental alterations,
and adverse life events including chronic or acute stress. If many
of these different etiological factors are expressed behaviorally
through final common neurophysiological features, identification of
these putative endophenotypes could not only provide a basis for
understanding of the disease but also enable rapid development of
novel selective classes of antidepressant treatments.
[0005] Candidate neural structures pertinent to depression
physiology have been identified in part by using structural and
functional imaging. Human fMRI studies have demonstrated altered
blood flow associated with depression in several brain regions,
including specific components of the emotion-regulating limbic
circuitry. In particular, the hippocampus has received considerable
attention as an integral component of the limbic system that
communicates directly with and drives other brain regions
implicated in depression, such as the prefrontal cortex,
hypothalamic-pituitary-adrenal (HPA) axis, and reward centers. A
substantial body of work favors the concept that the hippocampus is
hyperactive in depression.
[0006] PET imaging has been used in depressed patients to implicate
overactive excitatory pathways radiating from the hippocampus to
downstream cortical regions (e.g., to Cg25) which is thought to be
overactive in depression, and to orbitofrontal cortex), and
furthermore found that fluoxetine-induced reduction in hippocampal
activity was tightly linked to successful clinical response.
Meta-analysis of functional brain imaging in medication treatment
of depression indicated that changes in downstream cortical regions
are delayed until specific adaptive changes occur in the source of
primary afferent inputs, e.g., the hippocampus. This work showed
that the hippocampus is a "primary site of action" for major
antidepressants and a key initiator of successful response to
antidepressant treatment. Complicating this picture, however, is
evidence suggesting reduced hippocampal activity in depression,
including reduced hippocampal size in clinical depression, the fact
that excitatory hippocampal neurons display atrophy and death due
to chronic stress and stress hormone exposure, and the observation
that antidepressant-induced production of presumed excitatory
neurons in the dentate gyrus of the hippocampal formation is linked
to behavioral efficacy.
SUMMARY
[0007] Aspects of the present invention involve the implementation
of new optical technologies that allow sufficient spatial (pm) and
temporal resolution (ms) of electrical activity in distinct neural
circuits.
[0008] Consistent with one embodiment of the present invention, a
method is implemented for assessing neural activity in a neural
region having multiple subfields. An electrical response is evoked
in at least one subfield due to neural activity in the neural
region. Image data of the electrical response is captured at a
level sufficiently detailed in space and time to differentiate
between polarization-based events of two respective portions of the
subfield. Neural activity is assessed by correlating space and time
information, from the captured data, for the two respective
portions of the sub-field.
[0009] Consistent with one embodiment of the present invention, a
method is implemented for assessing neural activity in a neuronal
network that includes first and second portions electrically
related to one another. The neuronal network is stained with a
voltage sensitive dye. The first portion of the neuronal network is
stimulated. Responsive to the stimulation, image data is captured
that results from the voltage sensitive dye and neural activity in
the first portion. The image data is processed to assess neural
activity indicative of a disorder.
[0010] Consistent with one embodiment of the present invention, a
system is implemented for determining neural activity in a neural
region having multiple subfields. A preparation arrangement
prepares the neural region for imaging A stimulation arrangement
stimulates at least one subfield in the neural region. An imaging
device for captures image data resulting from stimulation of the
neural region, wherein the image data is captured at a level
sufficiently detailed in space and time to differentiate between
polarization-based events of two respective portions of the
subfield. A processor assesses neural activity by correlating space
and time information, from the captured data, for the two
respective portions of the sub-field.
[0011] Consistent with one embodiment of the present invention, an
arrangement is implemented for use in a system for determining
neural activity in a neural region having multiple subfields, the
system including a preparation arrangement for preparing the neural
region for imaging, a stimulation arrangement for stimulating at
least one subfield in the neural region, and an imaging device
capable of capturing image data resulting from stimulation of the
neural region. The arrangement includes a processor programmed and
adapted to process image data captured at a level sufficiently
detailed in space and time to differentiate between
polarization-based events of two respective portions of the
subfield, to assess neural activity by correlating space and time
information, from the captured data, for the two respective
portions of the sub-field.
[0012] Consistent with one embodiment of the present invention,
storage medium for use in a system for determining neural activity
in a neural region having multiple subfields, the system including
a preparation arrangement for preparing the neural region for
imaging, a stimulation arrangement for stimulating at least one
subfield in the neural region, and an imaging device capable of
capturing image data resulting from stimulation of the neural
region, a storage medium storing computer-executable data which,
when executed by a computer arrangement, cause the computer
arrangement to perform steps. A first step involves processing
image data captured at a level sufficiently detailed in space and
time to differentiate between polarization-based events of two
respective portions of the subfield. A second step involves
assessing neural activity by correlating space and time
information, from the captured data, for the two respective
portions of the sub-field.
[0013] The above summary of the present invention is not intended
to describe each illustrated embodiment or every implementation of
the present invention. The figures and detailed description that
follow more particularly exemplify these embodiments.
BRIEF DESCRIPTION OF DRAWINGS
[0014] The invention may be understood in consideration of the
detailed description of various embodiments of the invention that
follows in connection with the accompanying drawings in which:
[0015] FIG. 1A shows representative filmstrip acquired using
voltage-sensitive-dye imaging (VSDI), consistent with an example
embodiment of the present invention;
[0016] FIG. 1B shows pharmacological dissection of the VSDI signal,
consistent with an example embodiment of the present invention;
[0017] FIG. 1C shows single-pixel response (.DELTA.F/F vs. time,
post-averaging and filtering; top) from the indicated region to the
given stimulus train (bottom), consistent with an example
embodiment of the present invention;
[0018] FIG. 1D shows the results of cross-correlation analysis and
region of interest extraction, consistent with an example
embodiment of the present invention;
[0019] FIG. 2A shows results of a chronic mild stress (CMS) induced
ethologically relevant depressed-like state and animals treated
chronically with the indicated drug condition to model
antidepressant and antipsychotic treatment, consistent with an
example embodiment of the present invention;
[0020] FIG. 2B shows linear and quantitative response of the VSDI
total activity signal to applied stimulus current in both the
dentate gyms (DG) and Cornu Ammonis region CA1, consistent with an
example embodiment of the present invention;
[0021] FIG. 2C shows evoked DG activity was significantly decreased
in CMS-treated animals, consistent with an example embodiment of
the present invention;
[0022] FIG. 2D shows evoked CA1 activity was significantly
increased in CMS-treated animals, consistent with an example
embodiment of the present invention;
[0023] FIG. 2E shows relative activity between DG and CA1 activity,
calculated for each slice and averaged together for each animal,
consistent with an example embodiment of the present invention;
[0024] FIG. 2F shows evoked DG activity was significantly increased
in antidepressant, but not haloperidol, treated animals, consistent
with an example embodiment of the present invention;
[0025] FIG. 2G shows evoked CA1 activity was decreased in
antidepressant, but not haloperidol, treated animals, consistent
with an example embodiment of the present invention;
[0026] FIG. 2H shows relative activity between DG and CA1 activity
calculated for each slice and averaged together for each animal,
consistent with an example embodiment of the present invention;
[0027] FIG. 3A shows immobility times of CMS treated animals
relative to controls, and fluoxetine treatment in both control and
CMS groups, consistent with an example embodiment of the present
invention;
[0028] FIG. 3B shows activity of DG relative to CA1 in CMS-treated
animals and fluoxetine treated animals in both CMS and control
groups, consistent with an example embodiment of the present
invention;
[0029] FIG. 3C shows linear regression of the DG-CA1 relative
activity against the forced swim test (FST) scores for each
individual animal, consistent with an example embodiment of the
present invention;
[0030] FIG. 3D shows percent time in center on the open field test
(OFT) for treatment groups, consistent with an example embodiment
of the present invention;
[0031] FIG. 3E shows that no differences in total distance on the
OFT were observed for any treatment group indicating lack of
motility-related confounds, consistent with an example embodiment
of the present invention;
[0032] FIG. 3F shows that linear regression of the DG-CA1 relative
activity against the OFT scores for each individual animal,
consistent with an example embodiment of the present invention;
[0033] FIG. 4A shows unbiased stereological determination of BrdU+
cell density in the ventral hippocampus, consistent with an example
embodiment of the present invention;
[0034] FIG. 4B shows BrdU+ neuron density, consistent with an
example embodiment of the present invention;
[0035] FIG. 4C shows representative confocal images of the DG
labeled for BrdU, the mature neuronal marker NeuN, and the immature
neuronal marker Dcx, consistent with an example embodiment of the
present invention;
[0036] FIG. 5A shows a timeline for various steps in the imaging
process, consistent with an example embodiment of the present
invention;
[0037] FIG. 5B shows fluoxetine-treated animals immobility times,
consistent with an example embodiment of the present invention;
[0038] FIG. 5C shows DG-CA1 relative activity with fluoxetine
treatment, consistent with an example embodiment of the present
invention;
[0039] FIG. 5D shows the number of total BrdU+ cells per
hippocampus, consistent with an example embodiment of the present
invention;
[0040] FIG. 5E shows the number of BrdU+ cells, consistent with an
example embodiment of the present invention;
[0041] FIG. 5F shows the number of doublecortin (Dcx) cells,
consistent with an example embodiment of the present invention;
[0042] FIG. 5G shows representative confocal images of the DG
labeled for BrdU, the mature neuronal marker NeuN, and the
astrocytic marker GFAP, consistent with an example embodiment of
the present invention;
[0043] FIG. 6A shows a network model and resulting simulations,
consistent with an example embodiment of the present invention;
[0044] FIG. 6B shows results from network model and simulations,
consistent with an example embodiment of the present invention;
[0045] FIG. 6C shows data useful for assessing the experimental
effects of neurogenesis on the spread of activity in DG, consistent
with an example embodiment of the present invention;
[0046] FIG. 7 shows an example system for determining
characteristics of neuronal networks, according to an example
embodiment of the present invention;
[0047] FIG. 8 shows a flow diagram of a process for treatment of
depressed states in patients, according to an example embodiment of
the present invention;
[0048] FIG. 9A shows a wavelength diagram for use in a VSDI imaging
process, according to an example embodiment of the present
invention;
[0049] FIG. 9B shows an imaging apparatus for use in a VSDI imaging
process, according to an example embodiment of the present
invention;
[0050] FIG. 10A shows a sample VSDI frame and trace, processed
consistent with an example embodiment of the present invention;
[0051] FIG. 10B shows the trace indicated in FIG. 10A, imported
into MATLAB, consistent with an example embodiment of the present
invention;
[0052] FIG. 10C shows the 10 Hz stimulus profile used during
acquisition; consistent with an example embodiment of the present
invention;
[0053] FIG. 11A shows cross-comelogram for a single pixel produced
by cross-correlating the stimulus profile against the pixel
response, consistent with an example embodiment of the present
invention;
[0054] FIG. 11B shows Plot of maximal correlation amplitude and
phase of maximal amplitude for each pixel, consistent with an
example embodiment of the present invention;
[0055] FIG. 12A shows the result of plotting a calculation of
standard deviation of phases in the local surrounding area,
consistent with an example embodiment of the present invention;
[0056] FIG. 12B shows an initial, computationally extracted region
of interest, prior to smoothing and manual cropping, and also these
pixels overlaid on the phase and amplitude plots of FIG. 10B,
consistent with an example embodiment of the present invention;
[0057] FIG. 12C shows resulting images after morphological
smoothing and manual cropping, and also the result overlaid with
the phase and amplitude plots of FIG. 10B, consistent with an
example embodiment of the present invention;
[0058] FIG. 13A shows selective and specific effects of fluoxetine
and hippocampal physiology assessed by immunohistochemistry,
consistent with an example embodiment of the present invention;
[0059] FIG. 13B shows close-up images of CD68-positive activated
microglia in sham and irradiated hippocampi, consistent with an
example embodiment of the present invention;
[0060] FIG. 13C the effect of irradiation on hippocampal physiology
observed in the non-fluoxetine-treated animals, consistent with an
example embodiment of the present invention;
[0061] FIG. 13D shows the specificity of irradiation effects to
newborn neuron production in DG, consistent with an example
embodiment of the present invention;
[0062] FIG. 14A shows forced-swim test (FST) behaviors presented in
response to pharmacological treatment, expressed as a mean percent
time and consistent with an example embodiment of the present
invention;
[0063] FIG. 14B shows FST behaviors presented in response to CMS
and fluoxetine, expressed as a mean percent time and consistent
with an example embodiment of the present invention;
[0064] FIG. 14C shows FST behaviors presented in response to
irradiation and fluoxetine, expressed as a mean percent time and
consistent with an example embodiment of the present invention;
[0065] FIG. 15A shows a block diagram of a system for optical drug
screening, according to an example embodiment of the present
invention; and
[0066] FIG. 15B shows a specific system diagram of a large-format,
quasi-automated system for drug screening, according to an example
embodiment of the present invention.
[0067] While the invention is amenable to various modifications and
alternative forms, specifics thereof have been shown by way of
example in the drawings and will be described in detail. It should
be understood, however, that the intention is not to limit the
invention to the particular embodiments described. On the contrary,
the intention is to cover all modifications, equivalents, and
alternatives falling within the spirit and scope of the
invention.
DETAILED DESCRIPTION
[0068] The present invention is believed to be applicable to a
variety of different types of processes, devices and arrangements
involving imaging excitable tissue. Aspects of the present
invention have been found to be particularly advantageous in
applications benefitting from or involving assessment of neural
activity. While the present invention is not necessarily so
limited, various aspects of the invention may be appreciated
through a discussion of examples using this context.
[0069] Embodiments of the present invention are directed to
methods, systems and devices for assessing electrical activity of a
neural region. Aspects of the invention allow for assessment of
such electrical activity with high temporal and spatial precision.
High-speed and high-resolution imaging devices capture data
responsive to the electrical activity. Specialized processing
techniques are used to assess the electrical activity by
correlating time and phase information about the captured data.
[0070] In connection with various embodiments of the present
invention, a neural region is targeted for assessment of neural
activity. The neural region has subfields or local networks
consisting of a relatively small number of neural cell groups.
Image capture of a subfield is performed so as to provide
sufficient detail in space and time to differentiate between
polarization events (i.e., depolarization events or
hyperpolarization events) of respective portions of the subfield.
These portions can range from a single neuron to a small group of
neurons sufficient to distinguish individual polarization events
within the portions of the subfield. In a particular instance, the
respective portions represent one or more captured image pixels. In
one instance, the spatial detail is on the order of millimeters
(e.g., sub centimeter). In another instance, the spatial detail is
on the order of micrometers (e.g., sub millimeter). The temporal
detail can range from around a half second to a few milliseconds or
even faster.
[0071] One embodiment of the present invention involves the use of
spatial and temporal correlation techniques to assess the neural
activity of subfields. These techniques are designed to scale with
the precision of the image capture technology allowing for use with
a wide variety of current and future image capture
technologies.
[0072] Embodiments of the present invention use voltage sensitive
dye imaging (VSDI) to probe quantitatively the dynamics of other
neural networks following the induction, diagnosis or treatment of
medical states or diseases. One such embodiment uses VSDI to probe
quantitatively the differences in network activity due to various
factors including, but not limited to, those differences underlying
psychiatric disease and treatment.
[0073] Other embodiments of the present invention use other imaging
techniques, such as infrared imaging, near-infrared imaging or
optical topography. Such techniques involve the capture of image
data relative to the properties of blood within the neural region
(e.g., hemoglobin concentrations). Activity within the neural
region results in rapid changes to localized blood volume. The
captured image data includes data regarding the properties of the
blood within specific regions of the neural region. This data is
obtained through a determination of optical absorption
coefficients.
[0074] A specific embodiment of the present invention uses a VSDI
process to assess correlations between stimuli and resulting
electrical responses. Stimuli elicit responses from the neural
network. Voltage responsive dyes allow images of the responses to
be captured using an imaging device. Correlation techniques are
implemented to match the data from the imaging device to the
stimuli. For instance, delays are inherently present between the
application of the stimuli and the capturing of the image data.
Factors in the delay include, but are not limited to, propagation
delay through the neural network, delay in the voltage-sensitive
dye response, imaging device delays and processing delays. In a
particular embodiment, the stimuli profile is compared to the image
data with respect to time/phase. The comparison can be accomplished
by implementing a pixel-by-pixel correlation between the stimulus
and the pixel response. The pixel response can be determined as the
amount of change from one frame to the next.
[0075] One embodiment of the present invention uses voltage
sensitive dye imaging (VSDI) to probe quantitatively the dynamics
of neural networks (e.g., hippocampal) with millisecond resolution
following bidirectional affective state modulation (including both
induction and treatment of depressed-like states). It has been
found that a single measure of high-speed neurophysiological
activity-the evoked spread of activity through the dentate gyms of
the hippocampal formation, relative to that in CA1, accounts for
bidirectional changes in animal behavior in a manner that is
independent of the underlying mechanism of action of the affective
state modulators. These high-speed imaging results define a
network-level endophenotype for depression unifying disparate
findings in the literature, and demonstrate a tractable approach to
the understanding and treatment of neural substrates of psychiatric
disease.
[0076] While much of the discussion herein is directed to VSDI
and/or optical tomography, the invention is not so limited. To the
contrary the various embodiments, for example, the processing and
correlations techniques, are applicable to any number of different
data capture techniques. For instance, the processing and
correlation techniques can be applied to any data set that contains
temporal and spatial information having sufficient granularity and
precision.
[0077] In an example of one such an alternative embodiment, a slice
of brain is obtained which includes representative samples of the
prefrontal cortex, the dorsolateral prefrontal cortex, and the
subgenual cingulate. This block of tissue is treated with a calcium
dye. Various calcium dyes are able to resolve ion flux on a
cellular scale with microsecond precision. Using such dyes, the
activity of these regions can be used to assess the neural activity
of various regions and their interconnections. For example, an
assessment could be made as to whether the dorsolateral prefrontal
cortex has a direct correlation with the activity of the prefrontal
cortex, and inverse relationship with the activity of the subgenual
cingulate cells. In the case of depression, this may reach a steady
state in which the subgenual cingulate is relatively overactive as
compared with the prefrontal cortex. In the context of this
experimental paradigm, candidate drugs may be tried with respect to
their differential effect on these regions within the sample. Ideal
drugs for example might serve single targeted neurophysiological
roles, without creating additional neurophysiological effects which
are competitive to the therapeutic goal. For example, a drug which
succeeds in lowering activity levels in the subgenual cingulate
without lowering activity in the prefrontal cortex as well may be
preferable to one which attempts to lower activity generically in
both. Also in the context of such an embodiment, ideal locations
for neuromodulation of specific sites within the circuit may be
identified, with and without the benefit of synergistic
medications.
[0078] Readout or image capture of the physiological activity of
the neural network can be accomplished by a variety of techniques
including calcium imaging, biochemical imaging and infrared
imaging. PET and fMRI may also be applicable, albeit at lower
spatial and temporal resolutions.
[0079] Stimulating the circuit may be accomplished with a variety
of means that influence cellular activity, including application of
drugs, magnetic fields, electrical current, optical (including
opto-genetic) stimulation, ultrasound thermal and radiation methods
as are known in the art.
[0080] FIG. 7 shows an example system for determing characteristics
of neuronal networks, according to an example embodiment of the
present invention. Tissue preparation device 700 includes the
preparation of the neuronal network for various imaging techniques,
such as staining the tissue with voltage sensitive dye. In one
instance, tissue preparation device 700 includes a device to
stimulate the appropriate portion of the neuronal network. This can
be accomplished using electrodes, light (e.g., by modifying the
target network to react to light using optically responsive
channels or pumps) and/or similar techniques, as has been recently
published with reference to DBS applications. Image device 702
records the resulting signals. In one instance, image device 702 is
a high speed camera capable of capturing signals derived from
action potentials of the neuronal network. Imaging data 704 is
received from image device 702 by data processor 706. The data is
processed to filter unwanted signals, determine the target area and
to calculate the amount of activity. As discussed in more detail
herein, the process can include a number of steps including, but
not limited to, cross-correlation analysis between the stimulus and
the response, image smoothing and averaging, definition of the
region of interest, pixel to pixel cross correlation and
calculation of the mean amplitude of the region of interest.
[0081] Depending on the readout source, data may be captured for
example with a CCD camera, or as a digital matrix of sensor
readings obtained from a sensor grid, or serially positioned
sensors. Processing resultant readout data by correlating activity
level of cell types may be accomplished with computer software, for
image analysis applications such as Image J (U.S. National
Institutes of Health, Image J consortium). Providing processed
correlation results to the user may be provided through screen
displays, printed and transmitted data.
[0082] FIG. 8 shows a flow diagram of a process for treatment of
depressed states in patients, according to an example embodiment of
the present invention. At block 802 the depressed state is induced
on an animal, such as a rat. Alternatively, the depressed state
could be diagnosed rather than induced. This is particularly useful
for modifying this portion of the process to be used with human
patients. Treatments (e.g., drugs, dietary changes, environmental
changes) are administered to the depressed state animals at block
804. In one instance, the use of control groups and statistical
techniques can be used to bolster the validity of any results
obtained. At block 806, the amount of activity in the DG and CA1
areas of the brain are measured and quantified. The results of the
treatments can then be determined at block 808. This portion of the
process can then be repeated for different treatments and different
depressed states. Future patients can be diagnosed with a similar
depressed state at block 810. This can be accomplished through
various techniques, such as identifying the depressed state through
symptoms and by measuring the activity of the DG and CA1 areas of
the brain. Using the determination from block 808 the appropriate
treatment can then be administered.
[0083] The various methods and systems discussed herein (including
those discussed in connection with the flow diagram of FIG. 8) can
be applied to a number of applications other than those related to
either depression or the DG and CA1 subfields. A few example
disorders for which treatments can be screened include, but are not
limited to, Alzheimer's disease, mild cognitive impairment, autism,
bipolar disorder, schizophrenia, and Down Syndrome. Other instances
involve the use of the invention to assess areas of the human
nervous system including, but not limited to, the neocortex,
archicortex, paleocortex, cerebellum, medulla or spinal cord.
Within the hippocampus (a structure on the temporal lobe of the
neocortex) for example, pertinent cellular areas involved with a
specific circuit include the dentate gyms, the CA1 field, the CA3
fields. Various examples include, but are not limited to, the
upstream and downstream connections to the hippocampus (e.g., the
hypothalamus, frontal cortex, enthorhinal cortex, cingulate cortex,
mammillary bodies, septum, bed nucleus of stria terminalis,
amygdala, and nucleus accumbens). Other embodiments target
excitable cells other than those of the brain, such as the nervous
system or cardiac conduction system (e.g., sinoatrial node,
atrioventricular (AV) node, bundle of His; bundle branches;
internodal tracts, anterior-superior and posterior-inferior
divisions of the left bundle and the Purkinje network).
[0084] In order to assess the network activity a number of
different methods can be used to evoke a response in the target
neural region. A first example involves applying direct electrical
stimulation to a portion of the network. This can be accomplished
using, for example, patch clamping or other electrical probe
devices and methods. Another example involves the use of various
electromagnetic (EM) waves to evoke the desired response. Various
different EM wave sources are possible depending upon the type of
EM wave. A first example includes optical (visible or near-visible)
EM waves from a number of sources, such as light-emitting-diodes
(LEDs) or lasers. Another example of EM-type stimulus includes
gamma rays or X-rays. Additionally, a response can be evoked using
physical stimulus, such as physical perturbation provided by
ultra-sound. Ultra-sound can also be used to evoke a response by
heating the target neural region. Yet another mechanism for evoking
a response includes the use of a pharmacological or chemical agent
to affect the firing rate of the neural region.
[0085] For applications involving a live subject, in addition to
the stimulus discussed above, a response can be evoked by providing
any number of external stimuli to the subject including, but not
limited to, physical, mental and chemical stimuli.
[0086] A specific embodiment of the present invention can also be
implemented by relying upon intrinsic neural activity. In this
context, evoking a response involves providing conditions for the
neural network that are conducive to producing the intrinsic neural
activity. For example, the cardiac conduction system includes the
intrinsic firing of both the SA and AV nodes. Intrinsic network
activity from either or both of these nodes can be assessed by
providing proper conditions for the neural networks. Additionally,
the effects of various treatments (e.g., drugs or otherwise) on the
intrinsic network activity can also be assessed by first applying
the treatments and assessing the resulting network activity.
Another possible implementation for intrinsic neural activity is to
correlate interdependencies between areas within a subfield or
between different subfields.
[0087] FIG. 15A shows a basic block diagram of a system for
screening for neural network affecting drugs or treatments,
according to an embodiment of the invention. Optical control 104
communicates with database 102, optical source 106 and optical
detector 109. Optical source 106 provides optical stimulus to test
sample 108. Test sample 108 includes the drug under test, neural
cells that have optically responsive ion channels, and a
voltage/ion indicator. Alternatively test sample 108 does not
include a drug, but rather the neural cells have been treated in
some other manner.
[0088] Test sample 108 includes a neural network that is under
test. In a specific example, the neural network is representative
of one or more specific subfields of the brain. The effects of a
particular drug or treatment may be monitored for more than just
average activity over time. Using an embodiment of the present
invention, the effect of the treatment with respect to specific
activity patterns within a subfield can be monitored and
assessed.
[0089] In one instance, the indicator fluoresces in response to
light from optical source 106. Optical control 104 may also include
a reconfigurable readout, so that as different light-activated
membrane potential switches (LAMPS) and different light-emitting
indicators of cellular activity (LEIAs) are used, the same control
system can be readily adapted to each paradigm. Optical detector
109 produces a signal responsive to such florescence, and optical
control 104 receives the produced signal. The optical control 104
stores data obtained from the signal in database 102. The
information stored includes spatial and temporal data regarding the
electrical activity of the test samples 108. Other information may
include factors such as the intensity, duration and wavelength of
the detected light. In a particular instance, the stored data can
be compared against baseline data, where the baseline data
corresponds to data recorded prior to the introduction of the drug
to the test sample 108. In another instance, optical source 106 may
vary the intensity, duration or other parameters related to the
control of optical source 106. These and other parameters may be
stored in database 102. These parameters are then used to assess
the network activity.
[0090] It should be apparent that optical source 106 may be
implemented using a single light source, such as a light-emitting
diode (LED), or using several light sources. Similarly, optical
detector 109 may use one or more detectors and database 102 may be
implemented using any number of suitable storage devices.
[0091] FIG. 15B shows a system diagram of a large-format,
quasi-automated system for drug screening in accordance with a
specific embodiment of the invention. Control device 101 (e.g., a
computer or control logic) controls various processes, and serves
as the central point of system input/output functions. The
environment may be maintained at an appropriate temperature,
humidity, carbon dioxide level and ambient light level within the
walls of the climate control chamber 105, with the help of one or
more sensors 114 (e.g., thermostat, carbon dioxide sensor and
humidity sensor), carbon dioxide and humidifier apparatus 112, and
heater 110. Multi-well tray 141 contains test wells 140 for holding
cultured cells, drugs, and other ingredients needed for each test.
Tray 141 rests upon X-Y-Z table 125, the movement of which is
carried out by table actuators 120, under control of computer 101.
Xenon lamp 155 emits high-intensity white light 156, which is
passed through color filter 160. In the case that ChR2 is used for
stimulating the cells within wells 140, color filter 160 is blue,
causing blue light 161 to exit the filter, and strike dichroic
mirror 170. Blue light 161 then passes upward, through microscope
objective lens apparatus 130, and through the bottom of transparent
tray 141. In this fashion, the contents of wells 140, with their
transparent undersides, are illuminated. When a separate wavelength
of light is required to stimulate a fluorescent light-emitting
indicator of cellular activity, a filter of the appropriate
specification may be substituted for the previous filter 160,
causing light of the proper wavelength for this latter task to be
piped toward well 140. If the cells within well 140 have been
light-sensitized, and if the drug being tested in each of these
wells does not suppress the process, a light-emitting indicator of
cellular activity (LEIA), which has also been added to each well or
expressed by the cells via genetic modification, will emit light in
accordance with the voltage change caused by the effect of the
light. This second wavelength of light, which may be much smaller
in magnitude than the stimulation light, is collected by microscope
turret 135, will also be passed through dichroic mirror 175, onto
the lens of (CCD) camera 180.
[0092] Dichroic mirror 170 allows for upward reflection of both the
wavelength required to stimulate the optical gating of the membrane
(e.g., blue for ChR2), and the wavelength required by any LEIA used
(e.g., ultraviolet for FURA-2). This dichroic mirror may be
arranged to allow passage of the output spectrum of the LEIA (e.g.,
blue-green for FURA-2) with minimal reflection or absorption.
[0093] For further details regarding drug screening processes,
systems and devices, including those directed to a high-throughput
screening environment, such as array-based optical screening
system, reference can be made to U.S. patent application Ser. No.
60/996,116, to Zhang et al., filed on Aug. 10, 2007 and entitled
"Cell Line for Optically-Based Screening of Ion Channel
Modulators".
[0094] FIG. 9A shows a wavelength diagram for use in a VSDI imaging
process, according to an example embodiment of the present
invention. FIG. 9B shows an imaging apparatus for use in a VSDI
imaging process, according to an example embodiment of the present
invention. A specific embodiment of the present invention uses
light-activated channel/pump proteins, such as channel-rhodopsins
(ChR2) and halorhodopsins (NpHR), to stimulate the cell to be
imaged. The wavelength of light 902 that ChR2 responds strongly to
is around 470 nm. The wavelength of light 904 that NpHR responds
strongly to is around 590 nm. An example VSDI (RH-155) responds
strongly to light 906 having a wavelength of around 700 nm. FIG. 9B
shows that light 902 and 904 can be directed at target cells. The
difference in wavelengths is useful for allowing individual control
of the specific light-activated channel/pump proteins (e.g., one of
ChR2 or NpHR). A dichroic mirror can be used to direct light 902
and 904 toward the target cells. Light from the VSDI is captured by
a camera and processed accordingly. This all-optical
stimulation/inhibition/imaging process can be particularly useful
for, among other things, precise temporal and spatially controlled
stimulation and for avoiding disadvantages associated with
electrode-based stimulus or patch-clamp techniques. For further
details regarding such all-optical systems reference can be made to
"Integration of Light-Controlled Neuronal Firing and Fast Circuit
Imaging" by Airan, et al. (Current Opinion in Neurobiology 2007,
17:587-592), which is fully incorporated herein by reference.
[0095] The hippocampus is an example target for various embodiments
of the invention, as it is not only a central component of limbic
neural circuitry implicated in depression and drug response, but
also the gateway through which multimodal sensory information is
stored and flows to the limbic system. Models of hippocampal
function from the memory literature can be used as a framework for
understanding how the hippocampus may mediate affective responses.
These models assign fundamentally different roles to the
associative (e.g., dentate gyrus or DG) and output (CA1) subfields
of the hippocampal formation in spatial and temporal memory
processing. For example, the CA1 region is often conceptualized as
a comparator of signals received directly from the cortex, to
signals received from DG (by way of CA3), that then outputs this
difference or "error" signal to downstream cortical and subcortical
structures. The DG and CA1 local networks not only operate very
differently, but neurogenesis occurs only in the DG; moreover,
neuromodulators and stress hormones each can yield distinct effects
on the DG and CA1 subfields. In light of these differences, the
disparate clinical findings regarding hypoactivity or hyperactivity
of the hippocampal formation in depression could in principle be
reconciled with technology capable of separately resolving
electrical activity of the associative (e.g., DG) and output (CA1)
local circuits in behaviorally relevant paradigms. An example
application of VSDI imaging consistent with an embodiment of the
present invention is used to address such disparate clinical
findings. While the invention is not so limited, the discussion of
such a specific application can be useful to an understanding of
the present invention.
[0096] In probing the neurobiology of psychiatric disorders, rodent
models have proven useful in isolating key features that may
underlie disease etiology. Models of depression include paradigms
based on stress, learned helplessness, neurological lesion, and/or
genetic manipulation. In particular, the chronic mild stress (CMS)
paradigm has excellent predictive, face and constructs validities,
models core symptoms, and is considered to be an ethologically
relevant model of depression. Although molecular, synaptic,
cellular, and anatomic markers have been linked to depression or
antidepressants in several of these animal models of disease, a
neurophysiological endophenotype of depression has yet to be
identified that captures relevant changes in network activity with
high spatiotemporal resolution.
[0097] Applications of the present invention implement VSDI
technology as a powerful, quantitative tool to probe the
alterations in network activity to determine their contributions to
neuropsychiatric disease. For example, given the primary role of
the hippocampus in depression and its treatment, VSDI was used to
observe hippocampal activity directly with high spatial and
temporal resolution in rodent models of depression and
antidepressant treatment. The parameters of hippocampal activity
measured by VSDI reliably predict behavioral performance on a
forced swim test following combinations of CMS and chronic
antidepressant treatment. Specifically, CMS significantly reduces
activity in the DG and increases activity in CA1; moreover, chronic
antidepressant treatment induces opposite effects, confirming the
initial hypotheses. Together, these results identify the activity
of the DG relative to CA1 as a neurophysiological endophenotype of
depression that spans responses to distinct classes of
antidepressants, combinations of stress exposure and antidepressant
treatment, and multiple mechanisms of action on the cellular level.
The DG-CA1 relative activity, a subfield-resolved measure of
high-speed hippocampal electrical activity, unifies contradictory
and disparate findings in the field and may represent a common
pathway through which mechanistically diverse processes contribute
to depression and its treatment.
[0098] To use high-speed VSDI to explore neuronal network activity
changes in psychiatric disease models, experimental analysis
methods were developed to extract reliable quantitative features
from the imaging data across animals. A specific implementation of
one embodiment of the present invention was used to probe the
DG-CA1 activity in depressed mice. A description of the figures
related to this implementation follows, with specific details of
the implementation provided thereafter.
[0099] FIG. 1 shows voltage sensitive dye imaging (VSDI) of such
hippocampal network activity. FIG. 1A shows representative
filmstrip acquired using VSDI. Times indicated are relative to a
single pulse of stimulus current applied to the perforant path.
Warmer colors indicate greater activity (.DELTA.F/F) in the
corresponding pixel. Data represents the average of four individual
sweeps; traces were low-pass filtered (0.5 kHz cutoff), and frames
were spatially filtered with a 5.times.5 Gaussian spatial
average.
[0100] FIG. 1B shows pharmacological dissection of the VSDI signal.
The signal is initiated by excitatory synaptic transmission as it
is greatly reduced with NBQX application (10 .mu.M) and abolished
with concurrent D-AP5 application (25 .mu.M). GABAzine (2004) and
TTX (1 .mu.M) were applied subsequently to confirm signal
extinction.
[0101] FIG. 1C shows single-pixel response (.DELTA.F/F vs. time,
post-averaging and filtering; top) from the indicated region to the
given stimulus train (bottom).
[0102] FIG. 1D shows the results of cross-correlation analysis and
region of interest extraction. The phase (top left) and amplitude
(top right) of maximal correlation between the stimulus and
response at each pixel is also shown. The region responding to the
stimulus was extracted computationally based on similar phase
values of responding pixels (bottom).
[0103] FIG. 2 shows bi-directionally modified hippocampal network
dynamics in depressed-like states and antidepressant treatment.
FIG. 2A shows a chronic mild stress (CMS) induced ethologically
relevant depressed-like state (left). A separate group of animals
was treated chronically with the indicated drug condition to model
antidepressant and antipsychotic treatment (right, Ctrl--control,
FIX--fluoxetine, Imi--imipramine, Hal--haloperidol). After
completion of either protocol, a 5-minute forced swim test (FST)
was employed to probe behavior, and voltage sensitive dye imaging
(VSDI) employed to measure network activity in acute hippocampal
slices prepared from the same animals. All assays were performed
blind to treatment group and presented data are normalized by the
respective control means. CMS treated animals displayed
significantly increased immobility time relative to controls,
confirming depression-like behavior (left, student's t-test,
**p<0.01, n=6 animals per group). Fluoxetine and imipramine
animals show a decreased immobility time, confirming an
antidepressant effect, and no such effects were observed in
haloperidol treated animals (right, ANOVA, F.sub.3,22=29.46,
***p<0.001, n=5-6 animals per group).
[0104] FIG. 2B shows linear and quantitative response of the VSDI
total activity signal (mean signal amplitude times region of
interest area) to applied stimulus current in both the DG (top left
and middle, n=7 slices, R.sup.2=0.9855) and CA1 (bottom left and
right, n=5 slices, R.sup.2=0.9926). Left: Sample frames from
imaging of DG and CA1 responses, processed as in FIG. 1A.
[0105] FIG. 2C shows evoked DG activity was significantly decreased
in CMS-treated animals (student's t-test, *p<0.05, n=65
(control) and n=72 (CMS) evoked slice responses).
[0106] FIG. 2D shows evoked CA1 activity was significantly
increased in CMS-treated animals (left, student's t-test,
**p<0.01, n=79 (control) and n=75 (CMS) evoked slice
responses).
[0107] FIG. 2E shows relative activity between DG and CA1 activity,
calculated for each slice and averaged together for each animal.
Significant reduction was seen in CMS (student's t-test,
*p<0.05, n=6 animals per group).
[0108] FIG. 2F shows evoked DG activity was significantly increased
in antidepressant, but not haloperidol, treated animals (ANOVA,
F.sub.3,267=34.75, ***p<0.001, n=60-72 evoked slice responses
per group).
[0109] FIG. 2G shows evoked CA1 activity was decreased in
antidepressant, but not haloperidol, treated animals (ANOVA,
F.sub.3,270=5.173, **p<0.01, n=59-72 evoked slice responses per
group).
[0110] FIG. 2H shows relative activity between DG and CA1 activity
calculated for each slice and averaged together for each animal.
Significant increases were seen in both antidepressant treatment
groups, but absent in haloperidol group (ANOVA, F.sub.3,22=12.74,
**p<0.01, ***p<0.001), corresponding well to observed
behavior in FIG. 2A.
[0111] FIG. 3 shows that high speed network-dynamics measurements
correlate with antidepressant treatment of depressed-like
behavioral states. To develop the data depicted by FIG. 3, chronic
mild stress (CMS) was used for five weeks to induce a
depressed-like state and fluoxetine was administered during the
final two weeks of the stress protocol. After completion of this
protocol, a 5-minute forced swim test (FST) and open field test
(OFT) were employed to probe behavior, and voltage sensitive dye
imaging (VSDI) was employed to measure network activity in acute
hippocampal slices from the same animals. All assays were performed
blind to treatment group and FST and VSDI data are presented
normalized to the control mean.
[0112] FIG. 3A shows that CMS treated animals displayed
significantly increased immobility time relative to controls, and
fluoxetine treatment decreased immobility in both control and CMS
groups (ANOVA, F.sub.3,34=19.24, *p<0.05, **p<0.01,
***p<0.001, n=8-12 animals per group).
[0113] FIG. 3B shows that the activity of DG relative to CA1 was
significantly decreased in CMS-treated animals and fluoxetine
treatment significantly increased this statistic in both CMS and
control groups (ANOVA, F.sub.3,34=16.17, *p<0.05, ***
p<0.001, n=6-12 animals per group).
[0114] FIG. 3C shows linear regression of the DG-CA1 relative
activity against the FST scores for each individual animal
(r.sup.2=0.5545, p<10.sup.-6, n=35 individual animals)
demonstrating that the measured network activity accounts for more
than half of the variation in FST scores across groups.
[0115] FIG. 3D shows that no differences in percent time in center
on the OFT were observed for any treatment group indicating lack of
anxiety-related effects of treatment (ANOVA, F.sub.3,24=1.021,
P>0.05, n=4-9 animals per group).
[0116] FIG. 3E shows that no differences in total distance on the
OFT were observed for any treatment group indicating lack of
motility-related confounds (ANOVA, F.sub.3,20=1.776, P>0.05,
n=4-9 animals per group).
[0117] FIG. 3F shows that linear regression of the DG-CA1 relative
activity against the OFT scores for each individual animal
(r.sup.2=0.0306, p>0.4, n=25 individual animals) demonstrating
that the network activity measure does not correlate with
anxiety-related behavior.
[0118] FIG. 4 shows contrasting effects of antidepressant treatment
and chronic mild stress (CMS) on hippocampal neurogenesis.
[0119] FIG. 4A shows that unbiased stereological determination of
BrdU+ cell density in the ventral hippocampus from the same animals
represented in FIG. 3 shows an increased number of new cells with
fluoxetine treatment but no effect of CMS (ANOVA, F.sub.3,16=10.01,
n=3-5 animals per group, *p<0.05, **p<0.01). BrdU (50
mg/kg/day) was administered daily during the second week of
fluoxetine or vehicle.
[0120] FIG. 4B shows that phenotyping of BrdU+ cells for neuronal
markers in the ventral hippocampus results in an increased number
of new neurons with fluoxetine treatment but has no effect of CMS
(ANOVA, F.sub.3,15=20.37, n=3-5 animals per group,
***p<0.001).
[0121] FIG. 4C shows that representative confocal images of the DG
labeled for BrdU, the mature neuronal marker NeuN, and the immature
neuronal marker Dcx. Arrowheads indicate BrdU+ neurons.
[0122] FIG. 5 shows that neurogenesis is necessary for the network
dynamics and behavioral effects of antidepressant treatment.
[0123] FIG. 5A is a timeline showing various steps in the imaging
process. After an irradiation protocol (sham or 10 Gy/day over 2
days) designed to ablate hippocampal neurogenesis, female rats were
administered 20 mg/kg/day of fluoxetine or vehicle, followed by a 3
week delay for drug clearance and newborn neuron incorporation.
[0124] FIG. 5B shows that fluoxetine-treated animals showed a
decreased immobility time, confirming an antidepressant-like
effect, and no such effects were observed in irradiated animals
(ANOVA, F.sub.3,23=7.757, n=6 per group, *p<0.05,
**p<0.01).
[0125] FIG. 5C shows that significant increases were seen in DG-CA1
relative activity with fluoxetine treatment, but not following
irradiation (ANOVA, F.sub.3,22=3.997, n=5-6 animals per group,
*p<0.05). Results corresponded well to observed FST
behavior.
[0126] FIG. 5D shows that unbiased stereological determination of
total BrdU+ cells per hippocampus showed increased numbers of new
cells by weeklong fluoxetine treatment (20 mg/kg/day) and decreased
by 20 Gy irradiation (ANOVA, F.sub.3.24=48.92, n=4-8 animals per
group, **p<0.01, ***p<0.001). BrdU (50 mg/kg/day) was
administered concurrently with fluoxetine or vehicle.
[0127] FIG. 5E shows phenotyping of BrdU+ cells, determined by
scaling the total number of BrdU+ cells (from 5D) by the fraction
that expresses each fate-specific marker, revealed that fluoxetine
treatment specifically increased the total number of newborn
neurons (BrdU+/NeuN+) in the DG (student's t-test, p<0.05, n=6
animals per condition).
[0128] FIG. 5F shows immature neuronal marker doublecortin (Dcx)
staining was unaffected by fluoxetine, but was abolished by
irradiation (ANOVA, F.sub.3,24=94.44, n=4-8 animals per group,
***p<0.001).
[0129] FIG. 5G shows representative confocal images of the DG
labeled for BrdU, the mature neuronal marker NeuN, and the
astrocytic marker GFAP. Arrowheads indicate BrdU+ cells.
[0130] FIG. 6 shows mechanisms by which small numbers of network
elements can globally modulate network activity.
[0131] FIG. 6A Top: In this simple network model, a small fraction
of excitable elements were endowed with either a "lower threshold"
response to incoming activity (left), or implemented with
longer-range connections (right) creating a "small world" network.
Presented are results from simulations on networks with 0%, 6%,
12%, and 18% altered elements. Bottom: Images representing activity
increasing with the indicated fraction of altered elements
(averaged over 60 simulations), for each network model as above.
Images are contour maps of activity averaged across
simulations.
[0132] FIG. 6B shows quantification of results from both lowered
threshold (left) and small world (right) models. Top: Active area
(number of active elements) versus simulation time, for each
network model. Bottom: Maximum active area increases more steeply
than the number of altered elements. Summary graphs showing mean
(+SEM) of 60 simulations on networks constructed using the
indicated altered-element frequency.
[0133] FIG. 6C shows data useful for assessing the experimental
effects of neurogenesis on the spread of activity in DG. Signal
area and mean amplitude of recordings in this region were
separately analyzed for effects of treatments that modulated
neurogenesis.
[0134] While mean signal amplitude was not modulated by fluoxetine
or irradiation treatment (left; ANOVA, F.sub.3,179=2.390, n=37-60
evoked slice responses per group), the active network area was
modulated similarly to the total activity measure (right, ANOVA,
F.sub.3,180=5.502, n=37-60 evoked slice responses per group,
*p<0.05; FIG. 14C). All data was normalized by control mean.
[0135] FIG. 10A shows a screenshot taken from BV Analyzer
(Brainvision, RIKEN, Japan; the acquisition program used during
imaging) showing a sample VSDI frame and trace, processed
consistent with an example embodiment of the present invention.
FIG. 10B shows the trace indicated in FIG. 10A, imported into
MATLAB. FIG. 10C shows the 10 Hz stimulus profile used during this
acquisition.
[0136] FIG. 11A shows cross-comelogram for a single pixel produced
by cross-correlating the stimulus profile against the pixel
response, consistent with an example embodiment of the present
invention;
[0137] FIG. 11B shows a plot of maximal correlation amplitude
(right) and phase of maximal amplitude (left) for each pixel.
Correlation amplitude units are arbitrary and phase has units of
ms, with arbitrary zero point
[0138] FIG. 12A shows the result of plotting a calculation of
standard deviation of phases in the local surrounding area.
[0139] FIG. 12B shows an initial, computationally extracted region
of interest, prior to smoothing and manual cropping, and also these
pixels overlaid on the phase (middle) and amplitude (right) plots
of FIG. 10B.
[0140] FIG. 12C shows resulting images after morphological
smoothing and manual cropping, and also the result overlaid with
the phase (middle) and amplitude (right) plots of FIG. 10B.
[0141] FIG. 13A shows selective and specific effects of fluoxetine
and hippocampal physiology assessed by immunohistochemistry.
Irradiation effects on inflammatory cells were first assessed by
immunohistochemistry. Shown are images of labeled total microglia
(CDl lb) and activated microglia (CD68) in sham and irradiated
brains. Parallel immunohistochemistry for T cells, B cells, and
neutrophils had revealed no cells from these populations in the
brain tissue in either control or irradiated conditions. Top row:
Total CDllb population comparing typical sham and irradiated
brains; expected low levels of microglia are present in both
conditions. Middle row: Rare CD68-positive cells were seen in the
dentate gyrus of both sham and irradiated brains, with an
observable increase in the irradiated brains, as expected since the
microglial response is a major mechanism by which neurogenesis is
suppressed with irradiation. Bottom row: No CD68-positive cells
were observed in the cortex in either condition.
[0142] FIG. 13B shows close-up images of CD68-positive activated
microglia in sham (top) and irradiated (bottom) hippocampi.
[0143] FIG. 13C shows that despite the requisite increase in
activated microglia, no effect of irradiation on hippocampal
physiology was observed in the non-fluoxetine-treated animals. DG
VSDI revealed a significantly increased response in fluoxetine
treated animals, while irradiated animals showed no
fluoxetine-induced differences corresponding to the lack of new
neurons (ANOVA, F.sub.3,234=7.105, **p<0.01).
[0144] FIG. 13D further supports the specificity of irradiation
effects to newborn neuron production in DG. CA1 VSDI revealed a
significantly decreased response in fluoxetine treated animals that
was notably unaffected by irradiation (ANOVA, F.sub.3,278=4357,
*p<0.05). All data was normalized by control means. The lack of
irradiation effects in any control behavioral or physiological
assay confirms the specificity of this irradiation protocol for
ablating hippocampal neurogenesis under these conditions.
Additionally, the data reveal two effects of fluoxetine on the
hippocampus, a neurogenesis dependent increase of DG activity and a
neurogenesis-independent decrease of CA1 activity.
[0145] FIG. 14A shows FST behaviors presented in response to
pharmacological treatment, expressed as a mean percent time and
consistent with an example embodiment of the present invention.
Fluoxetine- and imipramine-treated animals show decreased
immobility times, and associated increases in swimming and
climbing, respectively, confirming an antidepressant effect. No
such effects were observed in haloperidol-treated animals (ANOVA,
F.sub.3,22=29.46 (mobility), F.sub.3,22=16.56 (swimming),
F.sub.3,22=18.64 (climbing), ***p<0.001, n=5-6 animals per
group).
[0146] FIG. 14B shows FST behaviors presented in response to CMS
and fluoxetine, expressed as a mean percent time and consistent
with an example embodiment of the present invention. CMS-treated
animals displayed significantly increased immobility time relative
to controls, and fluoxetine treatment decreased immobility in both
control and CMS groups. Fluoxetine and CMS+fluoxetine-treated
animals showed the associated increases in mobility behaviors
(swimming and climbing; ANOVA, F.sub.3,34=19.24 (immobility),
F.sub.3,34=9.964 (swimming), F.sub.3,34=5.675 (climbing),
*p<0.05, **p<0.01, ***p<0.001, n=8-12 animals per
group).
[0147] FIG. 14C shows FST behaviors presented in response to
irradiation and fluoxetine, expressed as a mean percent time and
consistent with an example embodiment of the present invention.
Fluoxetine animals showed decreases in immobility time and
increases in swimming, confirming an antidepressant effect, and no
such effects were observed in irradiated animals (ANOVA,
F.sub.3,23=7.757 (immobility), F.sub.3,23=5.904 (swimming),
F.sub.3,23=1.722 (climbing); *p<0.05, **p<0.01, n=6 animals
per group).
[0148] VSDI with a digital camera system optimized for high speed,
sensitivity, and resolution was used to observe the activity of
intact networks within acute horizontal slices prepared from the
ventral hippocampus of adult, female rats (FIG. 1A). Slices were
bath loaded with the fast-response voltage-sensitive dye
di-4-ANEPPS. After loading, slices were stimulated with
standardized electrical pulse trains and activity was recorded with
a 5 ms frame rate and 25 um spatial resolution. The signals
produced reflected total local network activity and were initiated
by excitatory transmission since the AMPA receptor blocker NBQX (10
.mu.M) nearly abolished the signal (FIG. 1B). Concurrent
application of the NMDA receptor blocker D-AP5 (25 .mu.M)
extinguished the signal as expected. To extract reliable
quantitative features of the network response from the high-speed
VSDI data, cross-correlation analysis was employed to take
advantage of the fact that the evoked responses occur with a
uniform fast time course and delay following stimulation and that
reverberatory oscillations or changes in response phase were never
observed under these conditions (FIGS. 1C, 1D, 10 and 11). This
cross-correlation analysis generated two plots, one corresponding
to the maximal response amplitude of each pixel (FIG. 1D, top
right, FIG. 11B, right), and the other corresponding to the time at
which this maximal amplitude occurred ("phase," FIG. 1D, top left,
FIG. 11B, left). Phase values were consistent in the regions
responding to stimulation, which were isolated computationally for
further analysis (FIG. 1D, bottom, FIG. 12). The area of this
region of interest and the mean amplitude of pixels in the region
were then computed, and local network response ("total activity")
was calculated as the mean signal amplitude multiplied by the area
of the region of interest. This measure was linear with the applied
stimulus current, making it a quantitative and reliable indicator
of network activity across animals (FIG. 2B).
[0149] Thus, the signal quality, consistency, and linear response
of the data showed the usefulness of this high-speed VSDI
technology for quantifying abnormalities of intact network dynamics
in CNS disease.
[0150] To explore hippocampal network activity changes induced in a
depression model, a 7-week chronic mild stress (CMS) paradigm was
used to elicit an ethologically relevant, depressed-like state in
adult, female rats (FIG. 2A, left). Behavioral responses were
measured with the modified Porsolt forced swim test (FST), the most
widely used animal test predictive of depressed-like states and
antidepressant responses. A model of chronic antidepressant
treatment was employed by using a 2-week dosing paradigm of
fluoxetine, an SSRI antidepressant, or imipramine, a tricylic
antidepressant (FIG. 2A, right). Haloperidol, a typical
antipsychotic, was separately administered to determine the
specificity of any observed effects to antidepressant treatment. In
each drug dosing protocol, a 48-hour delay was incorporated
following the last dose to permit behavioral testing in a drug-free
state. Experimenters were always blinded to treatment groups.
CMS-treated animals spent significantly more time immobile over a
5-minute FST compared to controls, indicating robust induction of a
depressed-like state (FIG. 2A, left). Moreover, animals treated
with either of the antidepressants showed significantly decreased
immobility on the FST, and this effect was absent in animals
treated with the antipsychotic haloperidol (FIG. 2A, right).
[0151] To determine the network activity changes associated with
these induced depression-related behavioral phenotypes, acute
hippocampal slices were generated from the same animals for
high-speed VSDI. Both the DG and CA1 networks were specifically
probed for the reasons described above, anticipating that there
would likely be different effects in each local network that might
nonetheless each be linked to the behavioral phenotype. For both of
these local networks, it was found that the total activity measure
is a reliable and quantitative indicator of evoked network activity
that is linear with the applied stimulus strength across animals
(FIG. 2B). In these and all experiments, the experimenters were
always blinded to treatment groups. Strikingly, DG activity was
significantly reduced in acute slices from CMS-treated animals
(FIG. 2C), while CA1 activity was significantly increased in slices
from these animals (FIG. 2D). The activity of the DG relative to
that of CA1 has consistently provided a reliable predictor of
depression-related behavioral phenotypes (FIG. 2E) that was robust
to experimental variation and allowed for quantitative comparisons
between individual animals
[0152] The CA1 aspect of this pattern is compatible with previous
work linking depression to elevated driving of pathways emerging
from the hippocampus, and the DG aspect is consistent with data
linking depression to reduced intra-hippocampal formation activity.
Together this data introduces the concept of an activity mismatch
between early associative and late output stages in hippocampal
processing. To validate this novel measure, the network-level
responses were probed in slices from the antidepressant
(fluoxetine, imipramine) or typical antipsychotic (haloperidol)
treated animals. Precisely the opposite pattern of network activity
was found in antidepressant-treated animals, which showed increased
activity in DG (FIG. 2F) and reduced activity in CA1 (FIG. 2G).
These effects were specific to the known mood-modulating agents, as
the typical antipsychotic haloperidol showed no effect on either
hippocampal network (FIG. 2F, 2G). Furthermore, the relative
activity of dentate to CA1 (FIG. 2H) continued to provide the most
reliable indicator of the behavioral phenotype on an
animal-to-animal basis (r.sup.2=0.5251, p<10.sup.-6, pooled
across CMS and antidepressant treatment groups).
[0153] To assess the generality of the results over conditions that
model the clinical use of antidepressants to treat depressed
states, animals were exposed to CMS for five weeks to induce a
depressed-like state and then administered fluoxetine chronically
during the last two weeks of the protocol to reverse the effects of
CMS (FIG. 3A). Like the previous antidepressant paradigm, a 48 hour
delay period was incorporated between the last fluoxetine dose and
the behavioral endpoint to enable testing in a drug free state, and
both behavioral and physiological experiments were conducted blind
to treatment condition.
[0154] Behavioral results confirmed the induction of a
depressed-like state in CMS-treated animals, and also the efficacy
of fluoxetine in reversing the behavioral effects of CMS (FIG. 3A).
Moreover, high-speed VSDI data showed that the relative activity
between DG and CA1 substantially accounted for these group
differences (FIG. 3B), and remarkably, on an individual animal
basis, this measure regressed linearly with the bidirectional
modulated FST scores and explained over half of the behavioral
variation (FIG. 3C). All four independent groups lie along the same
tightly correlated line, which would not be expected with random
variation within the groups and strongly supports the link between
the high-speed physiological metric and FST performance
(r.sup.2=0.5545, p<10.sup.-6 in this single experiment). Given
the large overlap in both FST and VSDI scores of these groups and
the sampling of data points across the full range of both these
axes, this circuit dynamics measure substantially accounts for
movement along the axis representing affective behavior.
[0155] In the same experiment, open field tests (OFT) were also
conducted to measure anxiety-related behavior. The relationship
between this anxiety measure and network dynamics responses was
investigated in order to determine the specificity of the
identified network phenotype for the depressed-like state
measurement of FST performance (FIG. 3D, 3E, 3F). In OFT, there
were no significant differences among the groups in the percent of
time the animals spent in the center of the field (FIG. 3D),
indicating that the behavioral manipulations were specific for
modulating correlates of the effects. Additionally, no effects were
seen on the total distance traveled by the animals (FIG. 3E),
indicating no confounding effects of the treatments on motility of
the animals. This data also indicates that the DG-CA1 relative
activity is specific for relating information on correlates of mood
as there was no correlation between this physiological metric and
the OFT scores (FIG. 3F; r.sup.2=0.0306, p>0.4).
[0156] Next changes in neurogenesis associated with these
bidirectional changes in behavior and network activity were probed
for, to determine if the network dynamics metric depends chiefly on
a single biological mechanism or instead retains its validity
across fundamentally different mechanisms that may underlie CMS and
antidepressant responses. Neurogenesis appears to be strongly
linked to antidepressant effects on the novelty-suppressed feeding
task, but has not been clearly implicated in depression-related
behavioral measures like the FST or in depression induction
involving chronic mild stress. However, acute and severe stress can
inhibit neurogenesis, suggesting that altered neurogenesis in some
settings could contribute to the etiology of depression. To
directly investigate whether the network and behavioral effects of
both CMS and antidepressant treatment could depend in part on a
common underlying neurogenesis mechanism, bromodeoxyuridine (BrdU)
was administered to label dividing cells during the last week of
treatment in the experiment described in connection with FIG. 3,
and blinded immunohistological counts from the hippocampus
contralateral to that used for VSDI were conducted with unbiased
stereology procedures, to determine the total number of newborn
cells (FIG. 4A, 4C). For this analysis the hippocampal formation
was divided into two components, dorsal and ventral. The ventral
hippocampus is thought to be more involved in mood regulation.
Thus, the acute slice VSDI experiments were conducted in ventral
hippocampus. In accord with previous observations, fluoxetine
administration robustly increased the newborn cell density in the
ventral hippocampus, both in the presence and absence of CMS
(control: 1.75 cells/mm.sup.3; CMS: 1.68 cells/mm.sup.3;
fluoxetine: 2.47 cells/mm.sup.3; CMS+fluoxetine: 2.38
cells/mm.sup.3; FIG. 4A); no such effects were observed in the
dorsal hippocampus (control: 3.21 cells/mm.sup.3; CMS: 3.09
cells/mm.sup.3; fluoxetine: 3.03 cells/mm.sup.3; CMS+fluoxetine:
3.41 cells/mm.sup.3; p>0.05 for all comparisons). Interestingly,
CMS did not significantly alter newborn cell density (FIG. 4A), in
either the dorsal or ventral hippocampus, suggesting that altered
neurogenesis is not present in this chronic mild stress protocol
and therefore may be very unlikely to mediate either the observed
depression-like behavior or the network-dynamics response of
CMS.
[0157] To definitively demonstrate changes or lack thereof in
neurogenesis, the BrdU+ cells were phenotyped using the mature
neuronal marker NeuN and the immature neuronal marker Doublecortin
(Dcx; FIGS. 4B, 4C). This analysis verified that fluoxetine
treatment specifically increased the density of NeuN and
Dcx-positive cells in the ventral BrdU+ population, and thus
neurogenesis, while the CMS protocol did not alter the fraction of
BrdU-positive cells expressing neuronal markers, confirming the
lack of effect of this CMS protocol on neurogenesis. Neurogenesis
changes might still be causative in some depressed states; this
experiment, carried out in the same animals from the experiment in
FIG. 3, which demonstrated robust effects of CMS on behavior and
network dynamics, indicates that depression-related interventions
can be linked to a final common pathway of altered high-speed
network dynamics even if initiated by fundamentally different
mechanisms.
[0158] To further probe the contribution of neurogenesis to the
depression-related behavioral and physiological effects of
antidepressant treatment, a novel protocol was developed to isolate
long-term effects of a temporally defined cohort of
fluoxetine-induced newborn neurons. This paradigm incorporates one
week of chronic fluoxetine treatment to trigger a burst of
neurogenesis, followed by a three week delay period to permit both
behavioral testing in a drug-free state and functional integration
of neurons born during antidepressant administration (the
newborn-neuron specific antidepressant protocol, NNS; FIG. 5A). To
control for non-neurogenesis-related effects of antidepressant
treatment, in some experimental groups hippocampal neurogenesis was
ablated via irradiation (10 Gy/day or sham treatment over two days)
one month prior to drug exposure; control experiments resolved no
physiological effects of irradiation alone on excitability/network
dynamics or behavior on the timescale of these experiments (FIG.
13). Under the NNS protocol, fluoxetine still significantly reduced
immobility on the FST (FIG. 5B). This effect was quenched by
irradiation, suggesting that hippocampal neurogenesis can be
required for depression-related effects of fluoxetine (FIG. 5B).
Correspondingly, high-speed VSDI data from the same animals
demonstrated that the NNS protocol also increased the activity of
DG relative to that of CA1, and that this effect relied similarly
on hippocampal neurogenesis (FIG. 5C). Notably, the increased DG
activity in response to fluoxetine treatment was significantly
dependent on neurogenesis while the decreased CA1 activity was
neurogenesis-independent (FIG. 5C); this result would be expected
since CA1 does not undergo neurogenesis, providing another control
for the specificity of the irradiation treatment under these
conditions (FIG. 13).
[0159] To quantify effects of the NNS antidepressant treatment on
the number of newborn cells and to confirm the selective reduction
of hippocampal neurogenesis by irradiation, BrdU was injected
during the week of drug treatment and conducted stereological
counts as described above. NNS-fluoxetine treatment increased the
BrdUi cell density as expected, and irradiation considerably
reduced the BrdU+ population (FIGS. 5D, 5G). Phenotyping the BrdUi
cells with NeuN and the glial marker GFAP confirmed the selective
increase in neurogenesis by the NNS-fluoxetine treatment (FIG. 5E).
Additionally, to assess the neurogenic environment at the time of
physiological analysis, the immature neuronal marker doublecortin
(Dcx) was employed. NNS-fluoxetine treatment did not affect the
number of Dcx+ cells, while irradiation eliminated Dcx+ cells (FIG.
5F). Since Dcx is only expressed for approximately 2 weeks after
last mitosis in newborn neurons, these data confirmed the permanent
effect of irradiation and the transient effect of NNS
antidepressant treatment.
[0160] These results suggest that while neurogenesis can be
important in eliciting behavioral and physiological effects of an
antidepressant, it does not follow that changes in neurogenesis
necessarily play a significant role in the etiology of the
depressed state. Behaviorally effective CMS treatment did not
decrease neurogenesis and ablation of hippocampal neurogenesis in
control animals did not induce a depressed-like state (FIGS. 4,
5B). Together, these findings support prior work implicating
neurogenesis in some behavioral tests modulated by antidepressants,
but indicate that the rate of neurogenesis does not by itself
reliably serve as a phenotype tracking behavioral performance, in
contrast to hippocampal network dynamics. The data demonstrate that
induction of a depressed-like state and antidepressant treatment
can share a common link to the network dynamics phenotype, without
requiring a common etiological relationship.
[0161] Although it has been shown that hippocampal neurogenesis has
an impact of antidepressant treatment on the DG network dynamics,
it is not immediately obvious how a modest increase in neurogenesis
could affect global network activity, especially with estimates
that newborn neurons normally comprise only several percent of the
DG population. To assess the ability of neurogenesis to implement
changes in physiological properties of the whole network, two
simple models were generated to study the effect of new neurons on
activity propagation (FIGS. 6A, 6B). These models were not
constructed to simulate the hippocampus explicitly, but to probe
important features of the observations that a small increase in the
number of new neurons can affect global flow of activity through
networks. In the first model the newborn neurons have a lowered
activity threshold, to model the increased excitability of newborn
neurons (FIG. 6A, top left). An alternate possibility is that the
newborn neurons simply wire differently; to address this
possibility, in the second model the new neurons adopt a distinct
connectivity pattern relative to the original neurons (FIG. 6A, top
right). Existing neurons were connected to only their nearest
neighbors while new neurons could generate longer range
connections, with lengths chosen from a power law distribution in
accord with measured neuronal projection distributions and wiring
cost perspectives. Such a network, in which a regular, local
connectivity pattern is perturbed by a few elements adopting longer
range connections, is known as a "small world" network and has been
shown to allow greater network synchronization and activity
propagation in more physiologically realistic neural network
models.
[0162] Dynamics of either model showed that networks with rare new
neurons could recruit greatly increased numbers of neurons into
activity, compared to networks with no new neurons (FIG. 6A, 6B).
Similar results were found in networks in which either afferent or
efferent synaptic efficiency was modulated to parallel the enhanced
synaptic plasticity associated with new neurons. These results
illustrate, over a range of biologically-plausible models, that
even small numbers of new elements can modulate global properties
of the network. It is important to note that these results also
underscore the importance of precise whole network imaging, as
slight differences in a small number of neurons can induce
non-intuitive, global changes in how information propagates through
neural networks without affecting local amplitudes of activity
(FIG. 6A). Since voltage sensitive dye imaging can quantify
activity changes of whole regions on the millisecond timescale, and
with micron spatial resolution, these results support VSDI use as a
powerful tool for determining how neural network activity is
altered in psychiatric disease.
[0163] Since the models demonstrate that adding small numbers of
neurons can profoundly affect the spatial extent of activity
spread, the experimental data was analyzed regarding features of
the experimental VSDI signal itself to determine the contribution
of altered signal area to the observed physiological effects of
neurogenesis in the DG. The total activity measure derived from the
VSDI signal is defined as the product of the area of the active
region and the mean amplitude of the signal. No appreciable effect
was found of antidepressant-induced neurogenesis on mean signal
amplitude (FIG. 5C, left). Instead, the effect of antidepressant
treatment on total activity could be explained mostly by an
increase in the spatial extent of the spread of activity through
the DG, which was blocked by irradiation (FIG. 5C, right). The
relative contribution of each signal component to total activity
was directly quantified by calculating the mutual information (a
measure of how much information one quantity contains about
another) between the total activity and either the active network
area or mean signal amplitude. With antidepressant treatment, the
mutual information between total activity and either area or mean
amplitude was noticeably different, with area yielding far more
information about total activity (0.415.+-.0.051 bits for area,
0.155.+-.0.038 bits for mean amplitude). This analysis demonstrates
that the neurogenesis-mediated change in the spatial extent of
network activity is a significant contributor to the modulation of
hippocampal activity observed in models of antidepressant
effect.
[0164] A combination of high-speed VSDI, structural analysis, and
behavioral testing were employed to probe for depression-linked
network abnormalities. First, fast VSD imaging and analysis
technology was developed to generate high spatiotemporal-resolution
maps of neuronal activity (FIG. 1); this methodology was
quantitative and displayed sufficient precision to allow direct
linkage of network dynamics to behavior across different treatment
groups (FIG. 2). Second, antidepressant and chronic mild stress
treatment were demonstrated to give rise to opposite effects on
activity in both the DG and CA1 regions of the hippocampal
formation, a finding which helps resolve disparate reports in the
depression literature (FIG. 2C, 2D, 2F, 2G). It then was shown that
the activity of the DG relative to that of CA1 defines a single
neurophysiological metric that substantially accounts for
mood-correlated behavioral phenotypes (FIG. 2, 3), and is specific
for antidepressant treatment medications (FIG. 2) and
depression-related behaviors (FIG. 3D, 3E, 3F). On an individual
animal basis this measure of hippocampal formation activity
regresses linearly with behavioral scores and accounts for over
half of the behavioral variation across groups (FIG. 3C), despite
fundamentally distinct mechanisms of action of antidepressant
treatment and depression-induction protocols (only the former
requiring a change in neurogenesis; FIG. 4, 5). This network-level
analysis allows for quantitative detection and mechanistic
investigation at the relevant spatial and temporal scales in models
of psychiatric disease. As substantial changes in network
properties were generated by changes in only a small number of
cells, and were reflected at the network-level in terms of activity
spread rather than amplitude (FIG. 6), it would have been virtually
impossible to detect these changes with typical field recording or
patch clamp methods. This underscores the unique importance of VSDI
in seeking network-level underpinnings of psychiatric disease.
[0165] It is important to consider how known etiologies of, and
treatments for, major depression are likely to modulate the DG-CA1
relative activity. Considering etiology first, depression-like
behavior can be induced without reducing the rate of neurogenesis
(FIG. 4), but stress and stress hormones can give rise to existing
cell loss in the DG with reduced dendritic arborization, and can
also give rise to hyperexcitable CA1 neurons, consistent with the
findings on both local networks.
[0166] Furthermore, it is certainly possible that acute and severe
stress, which can create vulnerability to depression, could reduce
new neuron production in human beings as in animals and impair
circuit plasticity. Regarding treatments, since insertion of new
neurons into the DG appears to be sufficient to drive relevant
network-dynamics changes, these data could account in part for the
therapeutic effects of the large number of antidepressant
treatments known to increase neurogenesis in animal models,
including SSRI's, TCAs, electroconvulsive therapy, lithium,
environmental enrichment, and exercise. While it is not formally
known if the antidepressant-induced newborn neurons in the DG are
excitatory or inhibitory themselves or which cell types their axons
target, addition of these neurons to the circuit appears to
increase total evoked DG network activity (FIGS. 4, 5). Moreover,
not only neuron production per se, but also antidepressant-induced
increases in neural plasticity and neuron survival could also give
rise to the same network dynamics effect, particularly if
conditions in the models are satisfied (e.g., the affected neurons
are more excitable or couple well to long-range excitatory
networks, such as the recurrent hilar mossy cell pathway (FIG.
6).
[0167] Some antidepressant treatments clearly do not directly
target the hippocampus, such as deep brain stimulation (DBS) which
can be targeted to Cg25 or accumbens. However, DBS is known to
reduce activity in Cg25 which receives important excitatory
projections from the hippocampus, suggesting that subgenual DBS can
intervene downstream of an overactive CA1. It had not been known
how to unify into a single model the hippocampal atrophy seen in
depression with the increase in excitatory driving of cortex from
hippocampus associated with clinical depression. These results
suggest that the increased activation in the subgenual cingulate
during depression could result in part from increased CA1 activity,
while reduced intrinsic hippocampal formation function in
depression would agree well with the decreased activity of the
DG.
[0168] In an interesting parallel to the observation that the ratio
of DG to CA1 activity in the ventral hippocampal formation predicts
mood-related behaviors, dorsal hippocampal memory storage models
have described roles for DG and CA1 that involve competitive and
comparative interactions in the two local networks. One class of
models suggests that CA1 functions in retrieval and transmission to
cortex of stored episodic memory, and that the DG gates this
retrieval based on contextual information. Other models suggest
that the DG, in combination with CA3, generates a predictive signal
that is sent to CA1, where this prediction is compared to
cortically driven signals representing sensory reality in order to
detect unexpected stimuli; the resulting difference signal is then
transmitted to cortex. Mood-related hippocampal dysfunction has not
been as well modeled but is thought to underlie aspects of the
cognitive symptomatology of depression including hopelessness,
which can manifest clinically as the inability to foresee or
navigate a reasonable and hopeful plan within the context of the
patient's environment. Antidepressant treatment that increases the
relative activity of the DG to CA1 could drive the hippocampal
comparative output in the direction of DG-derived predictive
signals, signaling an increase in the ability of the hippocampal
formation to model, predict, and plan for the future utilizing
contextual information. In contrast, decreasing the, activity of DG
relative to CA1, as observed in the depressed-like state, would
imply impaired recognition and predictive roles of the DG and an
inaccurate mismatch signal from CA1 driving cortical structures
like Cg25. Whether this signal is interpreted with negative valence
(signaling a poor model of the world and therefore hopelessness) or
positive valence would depend on concomitant reward pathway
activity involving distributed systems in the amygdala, accumbens
and mesolimbic dopamine projections.
[0169] Identification of this hippocampal neurophysiological
endophenotype serves as a starting point in mapping the
network-level changes in other brain regions implicated in
depression, such as the prefrontal and cingulate cortices,
amygdala, basal ganglia, and reward centers. While these other
brain regions are undoubtedly involved in depression physiology,
the ability of this hippocampal measure to convey information
regarding the animal's behavioral state supports the emerging
hypothesis that the hippocampus plays a primary role in mood
regulation, in addition to its accepted role in learning and
memory. High-speed, whole network analysis with VSDI is clearly
indicated to probe changes in other implicated brain regions, in
models of depression and other neuropsychiatric disorders. Using
the techniques and methods discussed herein it should be apparent
that more detailed models of altered activity flow are possible and
that this methodology can be extended to other depression models
and treatment paradigms. This putative endophenotype is
contemplated for use in screening for treatments that specifically
and similarly modulate hippocampal dynamics. Depression and
antidepressant-induced changes to neural circuitry can be monitored
using the techniques and systems discussed herein and may be useful
for achieving a quantitative understanding of the depressed
brain.
[0170] Accordingly, the various methods and systems described
herein can be used for the treatment of depressed states in
patients. In one such instance, the effectiveness of drugs and
other treatment techniques can be quantitatively measured and
tailored to minimize unwanted side-effects. In another instance,
the characteristics of the CA1 and DG in a patient can be used to
make diagnosis of the patient's depressed state. Whether
implemented in vitro on a neural network of a related specimen or
in vivo on the neural network of the same subject, these approaches
can be particularly useful for determining which patient treatments
would be most effective.
[0171] Such a method need not be conducted in slices. For example,
using infrared imaging, in which a high-resolution CCD camera is
implanted next to the brain structure of interest, the brain
circuits of a living organism can be scanned in real time, with
areas of interest analyzed in accordance with the presently
described method. As infrared tomography (from external to the
body) advances in capabilities, it is anticipated that these
methods will permit acquisition of data at sufficient temporal and
spatial resolution so as to serve within the context of the present
invention.
[0172] For further details regarding modeling depression through
use of various embodiments of the present invention reference can
be made to "High-Speed Imaging Reveals Neurophysiological Links to
Behavior in an Animal Model of Depression" by Airan et al.
(Science, Aug. 10, 2007, Vol. 317 pp. 819-823), which is fully
incorporated herein by reference.
[0173] While the VSDI signals received in embodiments of the
invention are robust and reliable, the known low signal to noise
ratio can be compensated for using various design and analyses. In
one such example, only evoked responses are considered so that
stimulus/response cross-correlation analysis can be utilized to
take advantage of the known response timing. Additionally, for
proper cross-slice and animal comparison, an algorithm is used to
automatically and efficiently extract relevant parameters of the
signal (region of interest, mean amplitude, etc.) from the data.
Peak response amplitude of -0.1% .DELTA.F/F was sufficient for
proper signal extraction.
[0174] The following algorithms provide an example implementation
for use in analyzing the VSDI data collected according to the
methods and systems described herein. To begin the analysis, the
raw VSDI data were imported into a software application, such as
MATLAB (Mathworks, Natick, Mass.). The data was represented as an
initial reference frame (F), followed by each imaging data frame,
which contained differential fluorescence values (.DELTA.F). For
each pixel and frame, the .DELTA.F/F value was calculated. Each
temporal frame was then spatially smoothed using a 3.times.3 pixel
digital Gaussian kernel (.sigma.=1):
1 16 .times. [ 1 2 1 2 4 2 1 2 1 ] ( 1 ) ##EQU00001##
[0175] During imaging, multiple individual sweeps (e.g., four
sweeps) are recorded of activity responding to stimulation with
each frequency and in each region. Following spatial averaging,
these four sweeps were averaged frame by frame, producing one VSDI
movie (FIG. 10). Next, each pixel in the movie was treated
independently, and a cross-correlation was computed.
.phi. sr [ m ] = n = - .infin. .infin. s [ n ] r [ n + m ] , ( 2 )
##EQU00002##
[0176] Where r[] is the pixel's response signal and s[] is the
stimulation profile used during acquisition (FIG. 11A).
[0177] The maximum of the cross-correlation amplitude
(max[O.sub.sr[m]]), which measures the system's response to
stimulation, was then found for each pixel, as well as the delay of
this peak (phase=arg[max[O.sub.sr[m]]). If there is an arbitrary
software delay of the camera system used, the phase values at this
stage are arbitrary with respect to the real latency of response to
stimulation.
[0178] Timing of the response with respect to the stimulus can be
calculated later using an absolute reference point inserted by
closing a shutter during acquisition prior to stimulation.
[0179] To extract the region of interest, it is observed that
pixels in this region tended to be clustered and of similar phase
to one other (FIG. 11B, left). In contrast, the phase is random and
incoherent in regions which might have high correlation amplitude,
but in reality correspond to noise or artifacts -compare the dark
red regions (non-tissue containing pixels) in FIG. 11B (right) to
their corresponding phases (left). To quantify this clustering, for
each pixel, the standard deviation (SD) of phase values in the
surrounding 7.times.7 pixel region was computed (FIG. 12A). Region
of interest pixels had a lower standard deviation, whereas noise
regions generate random phase values and correspondingly higher
standard deviation. The phases of the pixels that had the lowest SD
(threshold set at SD=32.5 ms in analysis) were calculated as the
initial distribution of phases in the region of interest.
[0180] To remove outlying values, the upper and lower 25% of the
distribution was replaced by a padding (initially plus or minus 25
ms). The pixels with phases in this range were defined to be part
of the region of interest. If the number of pixels in this region
was too low (less than 3% of the total frame size), this suggested
a relatively broad phase range of the true region of interest and
the padding was incremented until this number of pixels was above
an absolute threshold (3% of the frame size). The resultant pixels
formed the initial calculated region of interest (FIG. 12B).
[0181] Upon defining this initial region of interest, median
filtering with a 3.times.3 pixel window was completed to remove
spurious pixels outside the region of interest, and morphological
smoothing (first image-opening by 4 pixels and then image-closing
by 6 pixels) was completed to close gaps existing within the region
of interest. To remove any pixels outside the region of interest,
the resulting pixel maps were manually cropped with the user
blinded to treatment group producing the final, extracted region of
interest (FIG. 12C).
[0182] After extracting the region of interest, the mean
correlation amplitude of pixels within this region and the total
area of the region of interest (as a percentage of frame size) were
then calculated. The mean amplitude multiplied by this area ("total
activity") was the final statistic used to measure system response
for the voltage sensitive dye imaging data. Even though voltage
changes due to glia can be observed with voltage sensitive dyes,
the total activity measures reported here likely indicate only
neuronal responses to stimulation, as opposed to the slower
responses of hippocampal glia that are not appreciably sensitive to
AMPA and NMDA block especially given the short latency to peak of
the VSDI traces (6.3+/-0.6ms for 2 Hz stimulations), the lack of
summation of responses following 2 Hz stimulation, and the reliance
of the signal on AMPA and NMDA transmission (FIG. 1B).
[0183] In one embodiment of the present invention, the
aforementioned algorithms can be iteratively applied so as generate
a large-scale image of general neural activity. This large-scale
image can be more informative than less precise methods because the
image is derived from precise temporal and spatial data obtained
from a number of the relatively small portions of subfields.
[0184] Additionally, while asynchrony in the individual responses
in principle could yield a modulation of the correlation amplitude
relative to the absolute activity in the network, the phase values
local to each pixel were tightly correlated and not observably
patterned and no consistent effects on phase were noted between the
various treatments. These results combined with short latency of
the responses indicate that there is not a significant contribution
of asynchronous responses to extracellular stimulation in these
experiments. Furthermore, given the minimal role of the mean
amplitude in modulating total activity (FIG. 6C), it is unlikely
that asynchrony in the neuronal responses would significantly bias
this statistic. Accordingly, while modulation of the total activity
due to this asynchrony can not be ruled out, because of the reduced
role of mean in affecting total activity, the lack of variation of
the phase information between treatment groups, and the low
expected asynchrony in the evoked responses, the contribution of
this modulation is likely to be minimal to the observed effects on
network activity, which are quantified by total activity. The VSDI
signal amplitude represents a population signal similar to that of
a field-excitatory-postsynaptic potential (fEPSP) and the total
activity measure represents a signal similar to that of a fEPSP
integrated spatially across the region of interest and temporally
across the peak evoked responses to 10 stimulus pulses.
[0185] Aspects of the present invention can be used in combination
with a variety of neuron-directed applications, including those
which are discussed in the related background articles cited above
and listed herewith. For example, for detailed discussion of in
vivo VSDI applications, reference may be made to "Spatiotemporal
Dynamics of Sensory Responses in Layer 2/3 of Rat Barrel Cortex
Measured In Vivo by Voltage-Sensitive Dye Imaging Combined with
Whole-Cell Voltage Recordings and Neuron Reconstruction" (The
Journal of euroscience, Feb. 15, 2003, 23(3) pp. 1298-1309), which
is fully incorporated herein by reference. Another example that is
specifically directed to hippocampal CA1 aspects is discussed in
the article entitled "Hippocampal CA1 Circuitry Dynamically Gates
Direct Cortical Inputs Preferentially at Theta Frequencies" (The
Journal of Neuroscience, Oct. 19, 2005, 25(42) pp. 9567-9580),
which is fully incorporated herein by reference.
[0186] Various applications of the invention can be applied to any
disease where percolation of activity, excitation, inhibition, or
the ratio of excitation to inhibition between two or more cell
populations may be usefully addressed with imaging and processing
techniques disclosed herein. For example, in Alzheimer's disease,
the hippocampus and entorhinal cortex become diseased early. Mild
cognitive impairment (MCI) may be a hallmark of this early AD
process. Down Syndrome is manifested by a number of neurological
abnormalities developed as a result of gene expression from an
extra 21st chromosome. Among these are underdeveloped brain
structures including the temporal lobe, and a nearly universal
propensity for developing Alzheimer's disease. Autism,
schizophrenia, bipolar disorder and many other conditions may be
studied and improved treatments developed by identifying
circuit-level endophenotypes as described in the present
application, and using these as a well-circumscribed paradigm in
which to target candidate treatments.
[0187] Various embodiments of the invention are useful to elucidate
the role of one or more elements within a circuit. For example, one
aspect involves assessing the interdependency (or independence) of
areas of a subfield from the level of activity of areas of the
subfield circuit. Quantitative analysis allows interdependencies to
be identified and decoded.
[0188] Other embodiments of the invention can be used to elucidate
the location of the activity of a drug within that circuit. For
example, a drug may be added to the neural circuit, and the
locations of activity change (relative to activity without the
drug) are readout and recorded. These changes can be used to assess
both primary sites of action, and downstream effects upon the
greater circuit.
[0189] Another embodiment of the invention can be used to elucidate
the effect of a drug on a given circuit. For example, specifically
looking for fast acting/acute acting drugs, such as ketamine, for
depression, by bath applying the drugs to find the right acute drug
and drug dose that tunes activity in the desired manner in the
high-throughput screening setting, thereby identifying drugs that
act in days or hours rather than weeks/months
[0190] Other embodiments can be used to identify endophenotypes
that have the high predictive value for a given disorder or to
identify target locations for physical interventions including
electrical, magnetic stimulation and ultrasound and radiation
treatments.
[0191] The various embodiments described above are provided by way
of illustration only and should not be construed to limit the
invention. Based on the above discussion and illustrations, those
skilled in the art will readily recognize that various
modifications and changes may be made to the present invention
without strictly following the exemplary embodiments and
applications illustrated and described herein. For instance, such
changes may include uses for neurological characteristics other
than depression, such as other diseases, disorders and even the
study of normal activity. Other such changes include various in
vivo imaging implementations. Such modifications and changes do not
depart from the true spirit and scope of the present invention,
which is set forth in the following claims.
* * * * *