EEG/ECoG/EMG publications

Background Impaired neurological function affects 85% to 90% of cardiac arrest ( CA ) survivors. Pulsatile blood flow may play an important role in neurological recovery after CA . Cerebral blood flow ( CBF ) pulsatility immediately, during, and after CA and resuscitation has not been investigated. We characterized the effects of asphyxial CA on short‐term (<2 hours after CA ) CBF and femoral arterial blood pressure ( ABP ) pulsatility and studied their relationship to cerebrovascular resistance ( CVR ) and short‐term neuroelectrical recovery. Methods and Results Male rats underwent asphyxial CA followed by cardiopulmonary resuscitation. A multimodal platform combining laser speckle imaging, ABP , and electroencephalography to monitor CBF , peripheral blood pressure, and brain electrophysiology, respectively, was used. CBF and ABP pulsatility and CVR were assessed during baseline, CA , and multiple time points after resuscitation. Neuroelectrical recovery, a surrogate for neurological outcome, was assessed using quantitative electroencephalography 90 minutes after resuscitation. We found that CBF pulsatility differs significantly from baseline at all experimental time points with sustained deficits during the 2 hours of postresuscitation monitoring, whereas ABP pulsatility was relatively unaffected. Alterations in CBF pulsatility were inversely correlated with changes in CVR , but ABP pulsatility had no association to CVR . Interestingly, despite small changes in ABP pulsatility, higher ABP pulsatility was associated with worse neuroelectrical recovery, whereas CBF pulsatility had no association. Conclusions Our results reveal, for the first time, that CBF pulsatility and CVR are significantly altered in the short‐term postresuscitation period after CA . Nevertheless, higher ABP pulsatility appears to be inversely associated with neuroelectrical recovery, possibly caused by impaired cerebral autoregulation and/or more severe global cerebral ischemia.

Premotor areas of primates are specialized cortical regions that can contribute to hand movements by modulating the outputs of the primary motor cortex (M1). The goal of the present work was to study how the supplementary motor area (SMA) located within the same hemisphere [i.e., ipsilateral SMA (iSMA)] or the opposite hemisphere [i.e., contralateral (cSMA)] modulate the outputs of M1. We used paired-pulse protocols with intracortical stimulations in sedated capuchin monkeys. A conditioning stimulus in iSMA or cSMA was delivered simultaneously or before a test stimulus in M1 with different interstimulus intervals (ISIs) while electromyographic activity was recorded in hand and forearm muscles. The pattern of modulation from iSMA and cSMA shared some clear similarities. In particular, both areas predominantly induced facilitatory effects on M1 outputs with shorter ISIs and inhibitory effects with longer ISIs. However, the incidence and strength of facilitatory effects were greater for iSMA than cSMA. We then compared the pattern of modulatory effects from SMA to the ones from the dorsal and ventral premotor cortexes (PMd and PMv) collected in the same series of experiments. Among premotor areas, the impact of SMA on M1 outputs was always weaker than the one of either PMd or PMv, and this was regardless of the hemisphere, or the ISI, tested. These results show that SMA exerts a unique set of modulations on M1 outputs, which could support its specific function for the production of hand movements. NEW & NOTEWORTHY We unequivocally isolated stimulation to either the ipsilateral or contralateral supplementary motor area (SMA) using invasive techniques and compared their modulatory effects on the outputs of primary motor cortex (M1). Modulations from both SMAs shared many similarities. However, facilitatory effects evoked from ipsilateral SMA were more common and more powerful. This pattern differs from the ones of other premotor areas, which suggests that each premotor area makes unique contributions to the production of motor outputs.

Spike-and-wave discharges (SWDs) of absence epilepsy are considered as pathologic alterations of sleep spindles; however, their network-level relationship has never been convincingly revealed. In order to observe the development and generalization of the thalamocortical SWDs and the concomitant alterations of sleep related oscillations, we performed local field potential (LFP) and single unit recordings in rats for three months during their maturation. We found that while SWDs and spindles look similar in young, they become different with maturation and shift to appear in different brain states. Thus, despite being generated by the same network, they are likely two distinct manifestations of the thalamocortical activity. We show that while spindles are already mainly global oscillations, SWDs appear mainly only focally in young. They become capable to generalize later with maturation, when the out-of-focus brain regions develop a decreased inhibitory/excitatory balance. These results suggest that a hyperexcitable focus is not sufficient alone to drive generalized absence seizures. Importantly, we also found the gradual age dependent disappearance of sleep spindles coinciding with the simultaneous gradual emergence of spike and waves, which both could be reversed by the proper dosing of ethosuximide (ETX). Based on these observations we conclude that the absence seizure development might be a multi-step process, which might involve the functional impairment of cortical interneurons and network-level changes that negatively affect sleep quality.

The promotion of structural and functional plasticity by estrogens is a promising approach to enhance central nervous system function in the aged. However, how the sensitivity to estrogens is regulated across brain regions, age and experience is poorly understood. To ask if estradiol treatment impacts structural and functional plasticity in sensory cortices, we examined the acute effect of 17α-Estradiol in adult Long Evans rats following chronic monocular deprivation, a manipulation that reduces the strength and selectivity of deprived eye vision. Chronic monocular deprivation decreased thalamic input from the deprived eye to the binocular visual cortex and accelerated short-term depression of the deprived eye pathway, but did not change the density of excitatory synapses in primary visual cortex. Importantly, we found that the classical estrogen receptors ERα and ERβ were robustly expressed in the adult visual cortex, and that a single dose of 17α-Estradiol reduced the expression of the calcium-binding protein parvalbumin, decreased the integrity of the extracellular matrix and increased the size of excitatory postsynaptic densities. Furthermore, 17α-Estradiol enhanced experience-dependent plasticity in the amblyopic visual cortex, by promoting response potentiation of the pathway served by the non-deprived eye. The promotion of plasticity at synapses serving the non-deprived eye may reflect selectivity for synapses with an initially low probability of neurotransmitter release, and may inform strategies to remap spared inputs around a scotoma or a cortical infarct.

Spasmodic dysphonia (SD) is an incurable focal dystonia of the larynx that impairs speech and communication. Vibro-tactile stimulation (VTS) alters afferent proprioceptive input to sensorimotor cortex that controls speech. This proof-of-concept study examined the effect of laryngeal VTS on speech quality and cortical activity in 13 SD participants who vocalized the vowel /a/ while receiving VTS for 29 minutes. In response to VTS, 9 participants (69%) exhibited a reduction of voice breaks and/or a meaningful increase in smoothed cepstral peak prominence , an acoustic measure of voice/speech quality. Symptom improvements persisted for 20 minutes past VTS. Application of VTS induced a significant suppression of theta band power over the left somatosensory-motor cortex and a significant rise of gamma rhythm over right somatosensory-motor cortex. Such suppression of theta oscillations is observed in patients with cervical dystonia who apply effective sensory tricks, suggesting that VTS in SD may activate a similar neurophysiological mechanism. Results of this feasibility study indicate that laryngeal VTS modulates neuronal synchronization over sensorimotor cortex, which can induce short-term improvements in voice quality. The effects of long-term VTS and its optimal dosage for treating voice symptoms in SD are still unknown and require further systematic study.

Visual cortex (VC) over-activation analysed by evoked responses has been demonstrated in congenital deafness and after long-term acquired hearing loss in humans. However, permanent hearing deprivation has not yet been explored in animal models. Thus, the present study aimed to examine functional and molecular changes underlying the visual and auditory cross-modal reaction. For such purpose, we analysed cortical visual evoked potentials (VEPs) and the gene expression (RT-qPCR) of a set of markers for neuronal activation (c-Fos) and activity-dependent homeostatic compensation (Arc/Arg3.1). To determine the state of excitation and inhibition, we performed RT-qPCR and quantitative immunocytochemistry for excitatory (receptor subunits GluA2/3) and inhibitory (GABAA-α1, GABAB-R2, GAD65/67 and parvalbumin-PV) markers. VC over-activation was demonstrated by a significant increase in VEPs wave N1 and by up-regulation of the activity-dependent early genes c-Fos and Arc/Arg3.1 (thus confirming, by RT-qPCR, our previously published immunocytochemical results). GluA2 gene and protein expression were significantly increased in the auditory cortex (AC), particularly in layers 2/3 pyramidal neurons, but inhibitory markers (GAD65/67 and PV-GABA interneurons) were also significantly upregulated in the AC, indicating a concurrent increase in inhibition. Therefore, after permanent hearing loss in the rat, the VC is not only over-activated but also potentially balanced by homeostatic regulation, while excitatory and inhibitory markers remain imbalanced in the AC, most likely resulting from changes in horizontal intermodal regulation.

In audiovisual speech perception, visual information from a talker's face during mouth articulation is available before the onset of the corresponding audio speech, and thereby allows the perceiver to use visual information to predict the upcoming audio. This prediction from phonetically congruent visual information modulates audiovisual speech perception and leads to a decrease in N1 and P2 amplitudes and latencies compared to the perception of audio speech alone. Whether audiovisual experience, such as with musical training, influences this prediction is unclear, but if so, may explain some of the variations observed in previous research. The current study addresses whether audiovisual speech perception is affected by musical training, first assessing N1 and P2 event-related potentials (ERPs) and in addition, inter-trial phase coherence (ITPC). Musicians and non-musicians are presented the syllable, /ba/ in audio only (AO), video only (VO), and audiovisual (AV) conditions. With the predictory effect of mouth movement isolated from the AV speech (AV-VO), results showed that, compared to audio speech, both groups have a lower N1 latency and P2 amplitude and latency. Moreover, they also showed lower ITPCs in the delta, theta, and beta bands in audiovisual speech perception. However, musicians showed significant suppression of N1 amplitude and desynchronization in the alpha band in audiovisual speech, not present for non-musicians. Collectively, the current findings indicate that early sensory processing can be modified by musical experience, which in turn can explain some of the variations in previous AV speech perception research.

Gamma oscillations in visual cortex have been hypothesized to be critical for perception, cognition, and information transfer. However, observations of these oscillations in visual cortex vary widely; some studies report little to no stimulus-induced narrowband gamma oscillations, others report oscillations for only some stimuli, and yet others report large oscillations for most stimuli. To better understand this signal, we developed a model that predicts gamma responses for arbitrary images and validated this model on electrocorticography (ECoG) data from human visual cortex. The model computes variance across the outputs of spatially pooled orientation channels, and accurately predicts gamma amplitude across 86 images. Gamma responses were large for a small subset of stimuli, differing dramatically from fMRI and ECoG broadband (non-oscillatory) responses. We propose that gamma oscillations in visual cortex serve as a biomarker of gain control rather than being a fundamental mechanism for communicating visual information.

Animal behavioral paradigms, such as classical conditioning and operant conditioning, are an important tool to study the neural basis of cognition and behavior. These paradigms involve manipulating sensory stimuli in a way that learning processes are induced under controlled experimental conditions. However, the majority of the commercially available equipment did not offer flexibility to manipulate stimuli. Therefore, the development of most versatile devices would allow the study of more complex cognitive functions. The purpose of this work is to present a low-cost, customized and wireless-operated chamber for animal behavior conditioning, based on the joint operation of two microcontroller modules: Arduino Due and ESP8266-12E. Our results showed that the auditory stimulation system allows setting the carrier frequency in the range of 1 Hz up to more than 100 kHz and the sound stimulus can be modulated in amplitude, also over a wide range of frequencies. Likewise, foot-shock could be precisely manipulated regarding its amplitude (from ∼200 μA to ∼1500 μA) and frequency (up to 20 pulses per second). Finally, adult rats exposed to a protocol of cued fear conditioning in our device showed consistent behavioral response and electrophysiological evoked responses in the midbrain auditory pathway. Furthermore, the device developed in the current study represents an open source alternative to develop customized protocols to study fear memory under conditions of varied sensory stimuli.

Patients with Parkinson disease (PD) often experience nonmotor symptoms including cognitive deficits, depression, and anxiety. Cognitive and affective processes are thought to be mediated by prefrontal cortico-basal ganglia circuitry. However, the topography and neurophysiology of prefrontal cortical activity during complex tasks are not well characterized. We used high-resolution electrocorticography in pFC of patients with PD and essential tremor, during implantation of deep brain stimulator leads in the awake state, to understand disease-specific changes in prefrontal activity during an emotional face processing task. We found that patients with PD had less task-related theta–alpha power and greater task-related gamma power in the dorsolateral pFC, inferior frontal cortex, and lateral OFC. These findings support a model of prefrontal neurophysiological changes in the dopamine-depleted state, in which focal areas of hyperactivity in prefrontal cortical regions may compensate for impaired long-range interactions mediated by low-frequency rhythms. These distinct neurophysiological changes suggest that nonmotor circuits undergo characteristic changes in PD.

This work sought correlates of pitch perception, defined by neural activity above the lower limit of pitch (LLP), in auditory cortical neural ensembles, and examined their topographical distribution. Local field potentials (LFPs) were recorded in eight patients undergoing invasive recordings for pharmaco-resistant epilepsy. Stimuli consisted of bursts of broadband noise followed by regular interval noise (RIN). RIN was presented at rates below and above the LLP to distinguish responses related to the regularity of the stimulus and the presence of pitch itself. LFPs were recorded from human cortical homologues of auditory core, belt, and parabelt regions using multicontact depth electrodes implanted in Heschl's gyrus (HG) and Planum Temporale (PT), and subdural grid electrodes implanted over lateral superior temporal gyrus (STG). Evoked responses corresponding to the temporal regularity of the stimulus were assessed using autocorrelation of the evoked responses, and occurred for stimuli below and above the LLP. Induced responses throughout the high gamma range (60-200 Hz) were present for pitch values above the LLP, with onset latencies of approximately 70 ms. Mapping of the induced responses onto a common brain space demonstrated variability in the topographical distribution of high gamma responses across subjects. Induced responses were present throughout the length of HG and on PT, which is consistent with previous functional neuroimaging studies. Moreover, in each subject, a region within lateral STG showed robust induced responses at pitch-evoking stimulus rates. This work suggests a distributed representation of pitch processing in neural ensembles in human homologues of core and non-core auditory cortex.

Scientists find how syllable timing and stress are represented in human speech cortex.

The role of the claustrum in consciousness and vigilance states was proposed more than two decades ago; however, its role in anesthesia is not yet understood, and this requires more investigation. The aim of our study was to assess the impact of claustrum electrical stimulation during isoflurane anesthesia in adult rats. The claustrum in the left hemisphere was electrically stimulated using a bipolar tungsten electrode inserted stereotaxically. In order to monitor the anesthetic depth, the electrocorticogram (ECoG) was recorded before, during, and after claustrum stimulation using frontal and parietal epidural electrodes placed over the left hemisphere. After reaching stabilized slow-wave isoflurane anesthesia, twenty stimuli, each of one second duration with ten seconds interstimulus duration, were applied. ECoG analysis has shown that, after a delay from the beginning of stimulation, the slow-wave ECoG signal changed to a transient burst suppression (BS) pattern. Our results show that electrical stimulation of the claustrum area during slow-wave isoflurane anesthesia induces a transitory increase in anesthetic depth, documented by the appearance of a BS ECoG pattern, and suggests a potential role of claustrum in anesthesia.

Objective . Electrocorticography (ECoG) is commonly used to map epileptic foci and to implement brain–computer interfaces. Understanding the spatiotemporal correspondence between potentials recorded from the brain’s surface and the firing patterns of neurons within the cortex would inform the interpretation of ECoG signals and the design of (microfabricated) micro-ECoG electrode arrays. Based on the theory that synaptic potentials generated by neurons firing in synchrony superimpose to generate local field potentials (LFPs), we hypothesized that neurons in the cortex would fire at preferential phases of the micro-ECoG signal in a spatially dependent way. Approach . We custom fabricated micro-ECoG electrode arrays with a small opening for silicon arrays (NeuroNexus) to be inserted into the cortex. Main results . We found that the spectral coherence between micro-ECoG signals and intracortical LFPs decreased with distance and frequency, but the coherence with spiking units did not simply decrease over distance, likely due to the structure of the cortex. The majority of sorted units spiked during a preferred phase (usually downward) and frequency (usually below 20 Hz) of the micro-ECoG signal. Their preferred frequency decreased with administration of dexmeditomidine, a sedative commonly used for cortical mapping in patients with epilepsy prior to surgical resection. Dexmedetomidine concomitantly shifted the micro-ECoG spectral density towards lower frequencies. Therefore, the phase relationship between micro-ECoG signals and cortical spiking depends on the state of the brain, and spectrum shifts towards lower frequencies in the electrocorticography signal are a signature of increased spike-phase coupling. However, spike-phase coupling is not a static property since visual stimuli were found to modulate the magnitude of phase coupling at gamma frequency ranges (30–80 Hz), providing empirical evidence that neurons transiently phase-lock. Significance . The phase relationship between intracortical spikes and micro-ECoG signals depends on brain state, site separation, cortical structure, and external stimuli.

The studies described in this paper for the first time characterize the acute and chronic performance of optically transparent thin-film micro-electrocorticography (μECoG) grids implanted on a thinned skull as both an electrophysiological complement to existing thinned skull preparation for optical recordings/manipulations, and a less invasive alternative to epidural or subdurally placed μECoG arrays. In a longitudinal chronic study, μECoG grids placed on top of a thinned skull maintain impedances comparable to epidurally placed μECoG grids that are stable for periods of at least 1 month. Optogenetic activation of cortex is also reliably demonstrated through the optically transparent μECoG grids acutely placed on the thinned skull. Finally, spatially distinct electrophysiological recordings were evident on μECoG electrodes placed on a thinned skull separated by 500-750 μm, as assessed by stimulation evoked responses using optogenetic activation of cortex as well as invasive and epidermal stimulation of the sciatic and median nerve at chronic time points. Neural signals were collected through a thinned skull in mice and rats, demonstrating potential utility in neuroscience research applications such as in vivo imaging and optogenetics.

Spatial attention may be used to select target speech in one location while suppressing irrelevant speech in another. However, if perceptual resolution of spatial cues is weak, spatially focused attention may work poorly, leading to difficulty communicating in noisy settings. In electroencephalography (EEG), the distribution of alpha (8–14 Hz) power over parietal sensors reflects the spatial focus of attention [Banerjee, Snyder, Molholm, and Foxe (2011). J. Neurosci. 31, 9923–9932; Foxe and Snyder (2011). Front. Psychol. 2, 154.] If spatial attention is degraded, however, alpha may not be modulated across parietal sensors. A previously published behavioral and EEG study found that, compared to normal-hearing (NH) listeners, hearing-impaired (HI) listeners often had higher interaural time difference thresholds, worse performance when asked to report the content of an acoustic stream from a particular location, and weaker attentional modulation of neural responses evoked by sounds in a mixture [Dai, Best, and Shinn-Cunningham (2018). Proc. Natl. Acad. Sci. U. S. A. 115, E3286]. This study explored whether these same HI listeners also showed weaker alpha lateralization during the previously reported task. In NH listeners, hemispheric parietal alpha power was greater when the ipsilateral location was attended; this lateralization was stronger when competing melodies were separated by a larger spatial difference. In HI listeners, however, alpha was not lateralized across parietal sensors, consistent with a degraded ability to use spatial features to selectively attend.

The trigeminal nerve (cranial nerve V), along with other cranial nerves, has in recent years become a popular target for bioelectric medicine due to its direct access to neuromodulatory centers. Trigeminal nerve stimulation is currently being evaluated as an adjunctive treatment for various neurodegenerative and neuropsychiatric diseases despite the mechanism of action being in question. In this work, we describe the development and validation of a novel neural interface for the infraorbital branch of the trigeminal nerve utilizing a thin film (TF) nerve cuff containing multiple electrode sites allowing for more selective stimulation of the nerve. We characterized the properties of the device using electrochemical impedance spectroscopy, cyclic voltammetry, voltage excursions, and in vivo testing. Electrochemical measurements demonstrate that the platinum-based electrodes possess a capacitive charge carrying mechanism suitable for stimulation of biological tissue with a safe charge injection limit of 73.13 μC/cm2. In vivo stimulation experiments show that the TF cuff can reliably stimulate nerve targets eliciting cortical responses similar to a silicone cuff electrode. Furthermore, selecting different pairs of stimulation electrodes on the TF cuff modulated the magnitude and/or spatial pattern of cortical responses suggesting that the device may be able to selectively stimulate different parts of the nerve. These results suggest that the TF cuff is a viable neural interface for stimulation of the infraorbital branch of the trigeminal nerve that enables future research examining the therapeutic mechanisms of trigeminal nerve stimulation.

Spreading depolarization (SD) accompanies numerous neurological conditions, including migraine, stroke, and traumatic brain injury. There is significant interest in understanding the relationship between SD and neuronal injury. However, characteristics underlying SD and repolarization (RP) induced by global cerebral ischemia (e.g., cardiac arrest (CA)) and reperfusion are not well understood. Quantifying features of SD and RP during CA and cardiopulmonary resuscitation (CPR) may provide important metrics for diagnosis and prognosis of neurological injury from hypoxia-ischemia. We characterized SD and RP in a rodent model of asphyxial CA+CPR using a multimodal platform including electrocorticography (ECoG) and optical imaging. We detected SD and RP by (1) alternating current (AC), (2) direct current (DC), and (3) optical imaging of spreading ischemia, spreading edema, and vasoconstriction. Earlier SD (r=−0.80; p<0.001) and earlier RP (r=−0.71, p<0.001) were associated with better neurological recovery after 24hrs. SD+RP onset times predicted good vs poor neurological recovery with 82% sensitivity and 91% specificity. To our knowledge, this is the first preclinical study to link SD and RP characteristics with neurological recovery post-CA. These data suggest that SD and RP may be ultra-early, real-time prognostic markers of post-CA outcome, meriting further investigation into translational implications during global cerebral ischemia.

The functional organization of human auditory cortex can be probed by characterizing responses to various classes of sound at different anatomical locations. Along with histological studies this approach has revealed a primary field in posteromedial Heschl's gyrus (HG) with pronounced induced high-frequency (70–150 Hz) activity and short-latency responses that phase-lock to rapid transient sounds. Low-frequency neural oscillations are also relevant to stimulus processing and information flow, however, their distribution within auditory cortex has not been established. Alpha activity (7–14 Hz) in particular has been associated with processes that may differentially engage earlier versus later levels of the cortical hierarchy, including functional inhibition and the communication of sensory predictions. These theories derive largely from the study of occipitoparietal sources readily detectable in scalp electroencephalography. To characterize the anatomical basis and functional significance of less accessible temporal-lobe alpha activity we analyzed responses to sentences in seven human adults (4 female) with epilepsy who had been implanted with electrodes in superior temporal cortex. In contrast to primary cortex in posteromedial HG, a non-primary field in anterolateral HG was characterized by high spontaneous alpha activity that was strongly suppressed during auditory stimulation. Alpha-power suppression decreased with distance from anterolateral HG throughout superior temporal cortex, and was more pronounced for clear compared to degraded speech. This suppression could not be accounted for solely by a change in the slope of the power spectrum. The differential manifestation and stimulus-sensitivity of alpha oscillations across auditory fields should be accounted for in theories of their generation and function. SIGNIFICANCE STATEMENT To understand how auditory cortex is organized in support of perception, we recorded from patients implanted with electrodes for clinical reasons. This allowed measurement of activity in brain regions at different levels of sensory processing. Oscillations in the alpha range (7–14 Hz) have been associated with functions including sensory prediction and inhibition of regions handling irrelevant information, but their distribution within auditory cortex is not known. A key finding was that these oscillations dominated in one particular non-primary field, anterolateral Heschl's gyrus, and were suppressed when subjects listened to sentences. These results build on our knowledge of the functional organization of auditory cortex and provide anatomical constraints on theories of the generation and function of alpha oscillations.

A fundamental challenge in neuroscience is to understand what structure in the world is represented in spatially distributed patterns of neural activity from multiple single-trial measurements. This is often accomplished by learning a simple, linear transformations between neural features and features of the sensory stimuli or motor task. While successful in some early sensory processing areas, linear mappings are unlikely to be ideal tools for elucidating nonlinear, hierarchical representations of higher-order brain areas during complex tasks, such as the production of speech by humans. Here, we apply deep networks to predict produced speech syllables from a dataset of high gamma cortical surface electric potentials recorded from human sensorimotor cortex. We find that deep networks had higher decoding prediction accuracy compared to baseline models. Having established that deep networks extract more task relevant information from neural data sets relative to linear models (i.e., higher predictive accuracy), we next sought to demonstrate their utility as a data analysis tool for neuroscience. We first show that deep network's confusions revealed hierarchical latent structure in the neural data, which recapitulated the underlying articulatory nature of speech motor control. We next broadened the frequency features beyond high-gamma and identified a novel high-gamma-to-beta coupling during speech production. Finally, we used deep networks to compare task-relevant information in different neural frequency bands, and found that the high-gamma band contains the vast majority of information relevant for the speech prediction task, with little-to-no additional contribution from lower-frequency amplitudes. Together, these results demonstrate the utility of deep networks as a data analysis tool for basic and applied neuroscience.