Regularized matrix data clustering and its application to image analysis.

We propose a novel regularized mixture model for clustering matrix-valued data. The proposed method assumes a separable covariance structure for each cluster and imposes a sparsity structure (eg, low rankness, spatial sparsity) for the mean signal of each cluster. We formulate the problem as a finite mixture model of matrix-normal distributions with regularization terms, and then develop an expectation maximization type of algorithm for efficient computation. In theory, we show that the proposed estimators are strongly consistent for various choices of penalty functions. Simulation and two applications on brain signal studies confirm the excellent performance of the proposed method including a better prediction accuracy than the competitors and the scientific interpretability of the solution.

Modeling Spectral Properties in Stationary Processes of Varying Dimensions with Applications to Brain Local Field Potential Signals.

In some applications, it is important to compare the stochastic properties of two multivariate time series that have unequal dimensions. A new method is proposed to compare the spread of spectral information in two multivariate stationary processes with different dimensions. To measure discrepancies, a frequency specific spectral ratio (FS-ratio) statistic is proposed and its asymptotic properties are derived. The FS-ratio is blind to the dimension of the stationary process and captures the proportion of spectral power in various frequency bands. Here we develop a technique to automatically identify frequency bands that carry significant spectral power. We apply our method to track changes in the complexity of a 32-channel local field potential (LFP) signal from a rat following an experimentally induced stroke. At every epoch (a distinct time segment from the duration of the experiment), the nonstationary LFP signal is decomposed into stationary and nonstationary latent sources and the complexity is analyzed through these latent stationary sources and their dimensions that can change across epochs. The analysis indicates that spectral information in the Beta frequency band (12-30 Hertz) demonstrated the greatest change in structure and complexity due to the stroke.

Complete protection from impending stroke following permanent middle cerebral artery occlusion in awake, behaving rats.

Using a rodent model of ischemic stroke [permanent middle cerebral artery occlusion (pMCAO)], our laboratory has previously demonstrated that sensory-evoked cortical activation via mechanical single whisker stimulation treatment delivered under an anesthetized condition within 2 h of ischemic onset confers complete protection from impending infarct. There is a limited time window for this protection; rats that received the identical treatment at 3 h following ischemic onset lost neuronal function and sustained a substantial infarct. Rats in these studies, however, were anesthetized with sodium pentobarbital or isoflurane, whereas most human stroke patients are typically awake. To optimize our animal model, the present study examined, using functional imaging, histological, and behavioral analysis, whether self-induced sensorimotor stimulation is also protective in unrestrained, behaving rats that actively explore an enriched environment. Rats were revived from anesthesia either immediately or at 3 h after pMCAO, at which point they were allowed to freely explore an enriched environment. Rats that explored immediately after ischemic onset maintained normal cortical function and did not sustain infarct, even when their whiskers were clipped. Rats that were revived at 3 h post-pMCAO exhibited eliminated cortical function and sustained cortical infarct. Further, the data suggested that the level of individual active exploration could influence the outcome. Thus, early activation of the ischemic cortical area via unrestrained exploration resulted in protection from ischemic infarct, whereas late activation resulted in infarct, irrespective of the level of arousal or whisker-specific stimulation.

Early stimulation treatment provides complete sensory-induced protection from ischemic stroke under isoflurane anesthesia.

Using a rodent model of ischemia [permanent middle cerebral artery occlusion (pMCAO)], previous studies demonstrated that whisker stimulation treatment completely protects the cortex from impending stroke when initiated within 2 h following pMCAO. When initiated 3 h post-pMCAO, the identical treatment exacerbates stroke damage. Rats in these studies, however, were anesthetised with sodium pentobarbital, whereas human stroke patients are typically awake. To overcome this drawback, our laboratory has begun to use the anesthetic isoflurane, which allows rats to rapidly recover from pMCAO within minutes, to test stimulation treatment in awake rats and to determine whether isoflurane has an effect upon the pMCAO stroke model. We found no difference in infarct volume between pMCAO in untreated controls under either sodium pentobarbital or isoflurane, and the primary finding was that rats that received treatment immediately post-pMCAO maintain cortical function and no stroke damage, whereas rats that received treatment 3 h post-pMCAO exhibited eliminated cortical activity and extensive stroke damage. The only difference between anesthetics was the broad extent of evoked cortical activity observed during both functional imaging and electrophysiological recording, suggesting that the extent of evoked activity evident under isoflurane anesthesia is supported by underlying neuronal activity. Given the high degree of similarity with previous data, we conclude that the pMCAO stroke model is upheld with the use of isoflurane. This study demonstrated that the isoflurane-anesthetised rat pMCAO model can be used for cerebrovascular studies, and allows for highly detailed investigation of potential novel treatments for ischemic stroke using awake, behaving animals.

Astrocyte-neuron lactate shuttle plays a pivotal role in sensory-based neuroprotection in a rat model of permanent middle cerebral artery occlusion.

We have previously demonstrated protection from impending cortical ischemic stroke is achievable by sensory stimulation of the ischemic area in an adult rat model of permanent middle cerebral artery occlusion (pMCAo). We have further demonstrated that a major underpinning mechanism that is necessary for such protection is the system of collaterals among cerebral arteries that results in reperfusion of the MCA ischemic territory. However, since such collateral flow is weak, it may be necessary but not sufficient for protection and therefore we sought other complementary mechanisms that contribute to sensory-based protection. We hypothesized that astrocytes-neuron lactate shuttle (ANLS) activation could be another potential underpinning mechanism that complements collateral flow in the protection process. Supporting our hypothesis, using functional imaging, pharmacological treatments, and postmortem histology, we showed that ANLS played a pivotal role in sensory stimulation-based protection of cortex and therefore serves as the other supporting mechanism underpinning the protection process.

Astrocyte-neuron lactate shuttle plays a pivotal role in sensory-based neuroprotection in a rat model of permanent middle cerebral artery occlusion.

We have previously demonstrated protection from impending cortical stroke is achievable by sensory stimulation of the ischemic area in an adult rat model of permanent middle cerebral artery occlusion (pMCAo). We have further demonstrated that a major underpinning mechanism that is necessary for such protection is the system of collaterals among cerebral arteries that results in reperfusion of the MCA ischemic territory. However, since such collateral flow is weak, it may be necessary but not sufficient for protection and therefore we were seeking other complementary mechanisms that contribute to sensory-based protection. We hypothesized that astrocytes-to-neuron shuttle (ANLS) is another potential underpinning mechanism that could complement collateral flow in the protection process. Supporting our hypothesis, using functional imaging, pharmacological treatments, and postmortem histology, we show that ANLS has a pivotal role in sensory-based protection of cortex and therefor serves as the other supporting mechanism underpinning the protection process.

Development of highly sensitive, flexible dual L-glutamate and GABA microsensors for in vivo brain sensing.

Real-time tracking of neurotransmitter levels in vivo has been technically challenging due to the low spatiotemporal resolution of current methods. Since the imbalance of cortical excitation/inhibition (E:I) ratios are associated with a variety of neurological disorders, accurate monitoring of excitatory and inhibitory neurotransmitter levels is crucial for investigating the underlying neural mechanisms of these conditions. Specifically, levels of the excitatory neurotransmitter L-glutamate, and the inhibitory neurotransmitter GABA, are assumed to play critical roles in the E:I balance. Therefore, in this work, a flexible electrochemical microsensor is developed for real-time simultaneous detection of L-glutamate and GABA. The flexible polyimide substrate was used for easier handling during implantation and measurement, along with less brain damage. Further, by electrochemically depositing Pt-black nanostructures on the sensor's surface, the active surface area was enhanced for higher sensitivity. This dual neurotransmitter sensor probe was validated under various settings for its performance, including in vitro, ex vivo tests with glutamatergic neuronal cells and in vivo test with anesthetized rats. Additionally, the sensor's performance has been further investigated in terms of longevity and biocompatibility. Overall, our dual L-glutamate:GABA sensor microprobe has its unique features to enable accurate, real-time, and long-term monitoring of the E:I balance in vivo. Thus, this new tool should aid investigations of neural mechanisms of normal brain function and various neurological disorders.

Mild sensory stimulation reestablishes cortical function during the acute phase of ischemia.

When delivered within 1 and in most cases 2 h of permanent middle cerebral artery occlusion (pMCAO), mild sensory stimulation (intermittent single whisker stimulation) was shown to be completely neuroprotective 24 h after pMCAO in a rodent model of ischemic stroke, according to assessment with multiple techniques (Lay et al., 2010). The acute effect of stimulation treatment on the ischemic cortex, however, has yet to be reported. Here we characterize cortical function and perfusion during the 120 min whisker stimulation period in four experimental groups with treatment initiated 0, 1, 2 (protected groups), or 3 h (unprotected group) post-pMCAO using multiple techniques. According to functional imaging, a gradual return of evoked whisker functional representation to baseline levels was initiated with treatment onset and completed within the treatment period. Evoked neuronal activity and reperfusion to the ischemic area also showed a gradual recovery in protected animals. Surprisingly, a similar recovery profile was observed in response to treatment in all protected animals, regardless of treatment onset time. Nonstimulated pMCAO control group data demonstrate that reperfusion is not spontaneous. This makes the complete protection observed in the majority of animals stimulated at 2 h post-pMCAO even more surprising, as these animals recovered despite having been in a severely ischemic state for two full hours. In summary, when delivered within a 2 h window post-pMCAO, whisker stimulation treatment initiated reperfusion and a gradual recovery of cortical function that was completed or nearly completed within the treatment period.

Sensory stimulation-based protection from impending stroke following MCA occlusion is correlated with desynchronization of widespread spontaneous local field potentials.

In a rat model of ischemic stroke by permanent occlusion of the medial cerebral artery (pMCAo), we have demonstrated using continuous recordings by microelectrode array at the depth of the ischemic territory that there is an immediate wide-spread increase in spontaneous local field potential synchrony following pMCAo that was correlated with ischemic stroke damage, but such increase was not seen in control sham-surgery rats. We further found that the underpinning source of the synchrony increase is intermittent bursts of low multi-frequency oscillations. Here we show that such increase in spontaneous LFP synchrony after pMCAo can be reduced to pre-pMCAo baseline level by delivering early (immediately after pMCAo) protective sensory stimulation that reduced the underpinning bursts. However, the delivery of a late (3 h after pMCAo) destructive sensory stimulation had no influence on the elevated LFP synchrony and its underpinning bursts. Histology confirmed both protection for the early stimulation group and an infarct for the late stimulation group. These findings highlight the unexpected importance of spontaneous LFP and its synchrony as a predictive correlate of cerebral protection or stroke infarct during the hyperacute state following pMCAo and the potential clinical relevance of stimulation to reduce EEG synchrony in acute stroke.

Head Implants for the Neuroimaging of Awake, Head-Fixed Rats.

Anesthetics, commonly used in preclinical and fundamental scientific research, have a depressive influence on the metabolic, neuronal, and vascular functions of the brain and can adversely influence neurophysiological results. The use of awake animals for research studies is advantageous but poses the major challenge of keeping the animals calm and stationary to minimize motion artifacts throughout data acquisition. Awake imaging in smaller-sized rodents (e.g., mice) is very common but remains scant in rats as rats are bigger, stronger, and have a greater tendency to oppose movement restraints and head fixation over the long durations required for imaging. A new model of neuroimaging of awake, head-fixed rats using customized hand-sewn slings, 3D-printed head implants, head caps, and a headframe is described. The results acquired following a single trial of single-whisker stimulation suggest an increase in the intensity of the evoked functional response. The acquisition of the evoked functional response from awake, head-fixed rats is faster than that from anesthetized rats, reliable, reproducible, and can be used for repeated longitudinal studies.

Amount but not pattern of protective sensory stimulation alters recovery after permanent middle cerebral artery occlusion.

BACKGROUND AND PURPOSE: Using a rodent model of ischemia (permanent middle cerebral artery occlusion), our laboratory previously demonstrated that 4.27 minutes of patterned single-whisker stimulation delivered over 120 minutes can fully protect from impending damage when initiated within 2 hours of permanent middle cerebral artery occlusion ("early"). When initiated 3 hours postpermanent middle cerebral artery occlusion ("late"), stimulation resulted in irreversible damage. Here we investigate the effect of altering pattern, distribution, or amount of stimulation in this model. METHODS: We assessed the cortex using functional imaging and histological analysis with altered stimulation treatment protocols. In 2 groups of animals we administered the same number of whisker deflections but in a random rather than patterned fashion distributed either over 120 minutes or condensed into 10 minutes postpermanent middle cerebral artery occlusion. We also tested increased (full-whisker array versus single-whisker) stimulation. RESULTS: Early random whisker stimulation (condensed or dispersed) resulted in protection equivalent to early patterned stimulation. Early full-whisker array patterned stimulation also resulted in complete protection but promoted faster recovery. Late full-whisker array patterned stimulation, however, resulted in loss of evoked function and infarct volumes larger than those sustained by single-whisker counterparts. CONCLUSIONS: When induced early on after ischemic insult, stimulus-evoked cortical activity, irrespective of the parameters of peripheral stimulation that induced it, seems to be the important variable for neuroprotection.

Naturalistic experience transforms sensory maps in the adult cortex of caged animals.

Much of what is known about the functional organization and plasticity of adult sensory cortex is derived from animals housed in standard laboratory cages. Here we report that the transfer of adult rats reared in standard laboratory cages to a naturalistic habitat modifies the functional and morphological organization of the facial whisker representation in the somatosensory 'barrel' cortex. Cortical whisker representations, visualized with repeated intrinsic signal optical imaging in the same animals, contracted by 46% after four to six weeks of exposure to the naturalistic habitat. Acute, multi-site extracellular recordings demonstrated suppressed evoked neuronal responses and smaller, sharper constituent receptive fields in the upper cortical layers (II/III), but not in the thalamic recipient layer (IV), of rats with naturalistic experience. Morphological plasticity of the layer IV barrel field was observed, but on a substantially smaller scale than the functional plasticity. Thus, transferring animals to an environment that promotes the expression of natural, innate behaviours induces a large-scale functional refinement of cortical sensory maps.

Rapid development of strong, persistent, spatiotemporally extensive cortical synchrony and underlying oscillations following acute MCA focal ischemia.

Stroke is a leading cause of death and the leading cause of long-term disability, but its electrophysiological basis is poorly understood. Characterizing acute ischemic neuronal activity dynamics is important for understanding the temporal and spatial development of ischemic pathophysiology and determining neuronal activity signatures of ischemia. Using a 32-microelectrode array spanning the depth of cortex, electrophysiological recordings generated for the first time a continuous spatiotemporal profile of local field potentials (LFP) and multi-unit activity (MUA) before (baseline) and directly after (0-5 h) distal, permanent MCA occlusion (pMCAo) in a rat model. Although evoked activity persisted for hours after pMCAo with minor differences from baseline, spatiotemporal analyses of spontaneous activity revealed that LFP became spatially and temporally synchronized regardless of cortical depth within minutes after pMCAo and extended over large parts of cortex. Such enhanced post-ischemic synchrony was found to be driven by increased bursts of low multi-frequency oscillations and continued throughout the acute ischemic period whereas synchrony measures minimally changed over the same recording period in surgical sham controls. EEG recordings of a similar frequency range have been applied to successfully predict stroke damage and recovery, suggesting clear clinical relevance for our rat model.

Large-scale organization of rat sensorimotor cortex based on a motif of large activation spreads.

Parcellation according to function (e.g., visual, somatosensory, auditory, motor) is considered a fundamental property of sensorimotor cortical organization, traditionally defined from cytoarchitectonics and mapping studies relying on peak evoked neuronal activity. In the adult rat, stimulation of single whiskers evokes peak activity at topographically appropriate locations within somatosensory cortex and provides an example of cortical functional specificity. Here, we show that single whisker stimulation also evokes symmetrical areas of suprathreshold and subthreshold neuronal activation that spread extensively away from peak activity, effectively ignoring cortical borders by spilling deeply into multiple cortical territories of different modalities (auditory, visual and motor), where they were blocked by localized neuronal activity blocker injections and thus ruled out as possibly caused by "volume conductance." These symmetrical activity spreads were supported by underlying border-crossing, long-range horizontal connections as confirmed with transection experiments and injections of anterograde neuronal tracer experiments. We found such large evoked activation spreads and their underlying connections regardless of whisker identity, cortical layer, or axis of recorded responses, thereby revealing a large scale nonspecific organization of sensorimotor cortex based on a motif of large symmetrical activation spreads. Because the large activation spreads and their underlying horizontal connections ignore anatomical borders between cortical modalities, sensorimotor cortex could therefore be viewed as a continuous entity rather than a collection of discrete, delineated unimodal regions, an organization that could coexist with established specificity of cortical organization and that could serve as a substrate for associative learning, direct multimodal integration and recovery of function after injury.

Imaging cortical absorption, scattering, and hemodynamic response during ischemic stroke using spatially modulated near-infrared illumination.

We describe a technique that uses spatially modulated near-infrared (NIR) illumination to detect and map changes in both optical properties (absorption and reduced scattering parameters) and tissue composition (oxy- and deoxyhemoglobin, total hemoglobin, and oxygen saturation) during acute ischemic injury in the rat barrel cortex. Cerebral ischemia is induced using an open vascular occlusion technique of the middle cerebral artery (MCA). Diffuse reflected NIR light (680 to 980 nm) from the left parietal somatosensory cortex is detected by a CCD camera before and after MCA occlusion. Monte Carlo simulations are used to analyze the spatial frequency dependence of the reflected light to predict spatiotemporal changes in the distribution of tissue absorption and scattering properties in the brain. Experimental results from seven rats show a 17+/-4.7% increase in tissue concentration of deoxyhemoglobin and a 45+/-3.1, 23+/-5.4, and 21+/-2.2% decrease in oxyhemoglobin, total hemoglobin concentration and cerebral tissue oxygen saturation levels, respectively, 45 min following induction of cerebral ischemia. An ischemic index (I(isch)=ctHHbctO(2)Hb) reveals an average of more then twofold contrast after MCAo. The wavelength-dependence of the reduced scattering (i.e., scatter power) decreased by 35+/-10.3% after MCA occlusion. Compared to conventional CCD-based intrinsic signal optical imaging (ISOI), the use of structured illumination and model-based analysis allows for generation of separate maps of light absorption and scattering properties as well as tissue hemoglobin concentration. This potentially provides a powerful approach for quantitative monitoring and imaging of neurophysiology and metabolism with high spatiotemporal resolution.

Spatiotemporal dynamics of pial collateral blood flow following permanent middle cerebral artery occlusion in a rat model of sensory-based protection: a Doppler optical coherence tomography study.

There is a growing recognition regarding the importance of pial collateral flow in the protection from impending ischemic stroke both in preclinical and clinical studies. Collateral flow is also a major player in sensory stimulation-based protection from impending ischemic stroke. Doppler optical coherence tomography has been employed to image spatiotemporal patterns of collateral flow within the dorsal branches of the middle cerebral artery (MCA) as it provides a powerful tool for quantitative in vivo flow parameters imaging (velocity, flux, direction of flow, and radius of imaged branches). It was employed prior to and following dorsal permanent MCA occlusion (pMCAo) in rat models of treatment by protective sensory stimulation, untreated controls, or sham surgery controls. Unexpectedly, following pMCAo in the majority of subjects, some MCA branches continued to show anterograde blood flow patterns over time despite severing of the MCA. Further, in the presence of protective sensory stimulation, the anterograde velocity and flux were stronger and lasted longer than in retrograde flow branches, even within different branches of single subjects, but stimulated retrograde branches showed stronger flow parameters at 24 h. Our study suggests that the spatiotemporal patterns of collateral-based dorsal MCA flow are dynamic and provide a detailed description on the differential effects of protective sensory stimulation.

The triphasic intrinsic signal: implications for functional imaging.

Intrinsic signal optical imaging with red illumination (ISOI) is used extensively to provide high spatial resolution maps of stimulus-evoked hemodynamic-related signals as an indirect means to map evoked neuronal activity. This evoked signal is generally described as beginning with an undershoot or "dip" in signal that is faster, more transient, and weaker compared with the subsequent signal overshoot. In contrast, the evoked signal detected with blood oxygen level-dependent (BOLD) functional magnetic resonance imaging (fMRI) is generally described as containing an undershoot after the initial dip and overshoot, even though it, too, detects hemodynamic-related signals and its first two phases appear complementary to those of ISOI. Here, we used ISOI with 635 nm illumination to image over 13.5 s after a 1 s stimulus delivery to detect and successfully use the ISOI undershoot phase for functional mapping. Eight spatiotemporal attributes were assessed per signal phase including maximum areal extent and peak magnitude, both of which were largest for the ISOI overshoot, followed by the undershoot and then the initial dip. Peak activity location did not colocalize well between the three phases; furthermore, we found mostly modest correlations between attributes within each phase and sparse correlations between phases. Extended (13.5 s) electrophysiology recordings did not exhibit a reoccurrence of evoked suprathreshold or subthreshold neuronal responses that could be associated with the undershoot. Beyond the undershoot, additional overshoot/undershoot fluctuations were also mapped, but were typically less spatiotemporally specific to stimulus delivery. Implications for ISOI and BOLD fMRI are discussed.

Functional organization and plasticity in the adult rat barrel cortex: moving out-of-the-box.

Recent advances in functional imaging and neuronal recording techniques demonstrate that the spatial spread and amplitude of whisker functional representation in the somatosensory cortex of the adult rodent is extensive, but subject to modulations. One of the strongest modulators is naturalistic whisker use. In the cortices of rodents that have been transferred from their home cage to live for an extensive period in a naturalistic habitat, there is suppression of evoked neuronal responses accompanied by contraction and sharpening of receptive fields, and contraction and weakening of whisker functional representations. These unexpected characteristics also describe modulations of whisker functional representations in the cortex of a freely exploring rodent during short whisker-based explorations. These and related findings suggest that cortical modulations and plasticity could follow a 'less is more' strategy and, therefore, highlight how different cortical strategies could be utilized for different behavioral demands.

Long-Range, Border-Crossing, Horizontal Axon Radiations Are a Common Feature of Rat Neocortical Regions That Differ in Cytoarchitecture.

Employing wide-field optical imaging techniques supported by electrophysiological recordings, previous studies have demonstrated that stimulation of a spatially restricted area (point) in the sensory periphery results in a large evoked neuronal activity spread in mammalian primary cortices. In rats' primary cortices, such large evoked spreads extend diffusely in multiple directions, cross cortical cytoarchitectural borders and can trespass into other unimodal sensory areas. These point spreads are supported by a spatially matching, diffuse set of long-range horizontal projections within gray matter that extend in multiple directions and cross borders to interconnect different cortical areas. This horizontal projection system is in addition to well-known area-to-area clustered projections to defined targets through white matter. Could similar two-projection cortical systems also be found in cortical regions that differ in their cytoarchitectural structure? To address this question, an adeno-associated viral vector expressing green fluorescent protein (GFP) was injected as an anterograde tract tracer into granular somatosensory cortex (trunk area), dysgranular cortex (somatosensory dysgranular zone and extrastriate cortex) and agranular motor cortex (MCx). Irrespective of the injection site the same two projection systems were found, and their quantification revealed a close similarity to findings in primary sensory cortices. Following detailed reconstruction, the diffuse horizontal axon radiation was found to possess numerous varicosities and to include short, medium and long axons, the latter extending up to 5.2 mm. These "proof of concept" findings suggest that the similarity of the two projection systems among different cortical areas could potentially constitute a canonical motif of neocortical organization.

Hypertension prevents a sensory stimulation-based collateral therapeutic from protecting the cortex from impending ischemic stroke damage in a spontaneously hypersensitive rat model.

Assessing potential stroke treatments in the presence of risk factors can improve screening of treatments prior to clinical trials and is important in testing the efficacy of treatments in different patient populations. Here, we test our noninvasive, nonpharmacological sensory stimulation treatment in the presence of the main risk factor for ischemic stroke, hypertension. Utilizing functional imaging, blood flow imaging, and histology, we assessed spontaneously hypertensive rats (SHRs) pre- and post-permanent middle cerebral artery occlusion (pMCAO). Experimental groups included a treatment SHR group (sensory-stimulated group), control untreated SHR group (no sensory stimulation), and a treated (sensory-stimulated) Wistar-Kyoto normotensive group. Unlike our previous studies, which showed sensory-based complete protection from impending ischemic cortical stroke damage in rats as seen in the treated Wistar-Kyoto group, we found that SHRs at 24hr post-pMCAO lacked evoked cortical activation, had a significant reduction in blood flow within the MCA, and sustained very large infarcts regardless of whether they received stimulation treatment. If translatable, this work highlights a potential need for a combined treatment plan when delivering sensory stimulation treatment in this patient population.