Incorporating intrinsic suppression in deep neural networks captures dynamics of adaptation in neurophysiology and perception

Adaptation is a fundamental property of sensory systems that can change subjective experiences in the context of recent information. Adaptation has been postulated to arise from recurrent circuit mechanisms or as a consequence of neuronally intrinsic suppression. However, it is unclear whether intrinsic suppression by itself can account for effects beyond reduced responses. Here, we test the hypothesis that complex adaptation phenomena can emerge from intrinsic suppression cascading through a feedforward model of visual processing. A deep convolutional neural network with intrinsic suppression captured neural signatures of adaptation including novelty detection, enhancement, and tuning curve shifts, while producing aftereffects consistent with human perception. When adaptation was trained in a task where repeated input affects recognition performance, an intrinsic mechanism generalized better than a recurrent neural network. Our results demonstrate that feedforward propagation of intrinsic suppression changes the functional state of the network, reproducing key neurophysiological and perceptual properties of adaptation.

Automating Interictal Spike Detection: Revisiting A Simple Threshold Rule

Interictal spikes (IIS) are bursts of neuronal depolarization observed electrographically between periods of seizure activity in epilepsy patients. However, IISs are difficult to characterize morphologically and their effects on neurophysiology and cognitive function are poorly understood. Currently, IIS detection requires laborious manual assessment and marking of electroencephalography (EEG/iEEG) data. This practice is also subjective as the clinician has to select the mental threshold that EEG activity must exceed in order to be considered a spike. The work presented here details the development and implementation of a simple automated IIS detection algorithm. This preliminary study utilized intracranial EEG recordings collected from 7 epilepsy patients, and IISs were marked by a single physician for a total of 1339 IISs across 68 active electrodes. The proposed algorithm implements a simple threshold rule that scans through iEEG data and identifies IISs using various normalization techniques that eliminate the need for a more complex detector. The efficacy of the algorithm was determined by evaluating the sensitivity and specificity of the detector across a range of thresholds, and an approximate optimal threshold was determined using these results. With an average true positive rate of over 98% and a false positive rate of below 2%, the accuracy of this algorithm speaks to its use as a reliable diagnostic tool to detect IISs, which has direct applications in localizing where seizures start, detecting when seizures start, and in understanding cognitive impairment due to IISs. Furthermore, due to its speed and simplicity, this algorithm can be used for real-time detection of IIS that will ultimately allow physicians to study their clinical implications with high temporal resolution and individual adaptation.

Sparse but not 'grandmother-cell' coding in the medial temporal lobe

Although a large number of neuropsychological and imaging studies have demonstrated that the medial temporal lobe (MTL) plays an important role in human memory, there are few data regarding the activity of neurons involved in this process. The MTL receives massive inputs from visual cortical areas, and evidence over the last decade has consistently shown that MTL neurons respond selectively to complex visual stimuli. Here, we focus on how the activity patterns of these cells might reflect the transformation of visual percepts into long-term memories. Given the very sparse and abstract representation of visual information by these neurons, they could in principle be considered as 'grandmother cells'. However, we give several arguments that make such an extreme interpretation unlikely.

Invariant visual representation by single neurons in the human brain

It takes a fraction of a second to recognize a person or an object even when seen under strikingly different conditions. How such a robust, high-level representation is achieved by neurons in the human brain is still unclear. In monkeys, neurons in the upper stages of the ventral visual pathway respond to complex images such as faces and objects and show some degree of invariance to metric properties such as the stimulus size, position and viewing angle. We have previously shown that neurons in the human medial temporal lobe (MTL) fire selectively to images of faces, animals, objects or scenes. Here we report on a remarkable subset of MTL neurons that are selectively activated by strikingly different pictures of given individuals, landmarks or objects and in some cases even by letter strings with their names. These results suggest an invariant, sparse and explicit code, which might be important in the transformation of complex visual percepts into long-term and more abstract memories.

Amygdala-enriched genes identified by microarray technology are restricted to specific amygdaloid subnuclei

Microarray technology represents a potentially powerful method for identifying cell type- and regionally restricted genes expressed in the brain. Here we have combined a microarray analysis of differential gene expression among five selected brain regions, including the amygdala, cerebellum, hippocampus, olfactory bulb, and periaqueductal gray, with in situ hybridization. On average, 0.3% of the 34,000 genes interrogated were highly enriched in each of the five regions, relative to the others. In situ hybridization performed on a subset of amygdala-enriched genes confirmed in most cases the overall region-specificity predicted by the microarray data and identified additional sites of brain expression not examined on the microarrays. Strikingly, the majority of these genes exhibited boundaries of expression within the amygdala corresponding to cytoarchitectonically defined subnuclei. These results define a unique set of molecular markers for amygdaloid subnuclei and provide tools to genetically dissect their functional roles in different emotional behaviors.

Imagery neurons in the human brain

Vivid visual images can be voluntarily generated in our minds in the absence of simultaneous visual input. While trying to count the number of flowers in Van Gogh's Sunflowers, understanding a description or recalling a path, subjects report forming an image in their "mind's eye". Whether this process is accomplished by the same neuronal mechanisms as visual perception has long been a matter of debate. Evidence from functional imaging, psychophysics, neurological studies and monkey electrophysiology suggests a common process, yet there are patients with deficits in one but not the other. Here we directly investigated the neuronal substrates of visual recall by recording from single neurons in the human medial temporal lobe while the subjects were asked to imagine previously viewed images. We found single neurons in the hippocampus, amygdala, entorhinal cortex and parahippocampal gyrus that selectively altered their firing rates depending on the stimulus the subjects were imagining. Of the neurons that fired selectively during both vision and imagery, the majority (88%) had identical selectivity. Our study reveals single neuron correlates of volitional visual imagery in humans and suggests a common substrate for the processing of incoming visual information and visual recall.

Category-specific visual responses of single neurons in the human medial temporal lobe

The hippocampus, amygdala and entorhinal cortex receive convergent input from temporal neocortical regions specialized for processing complex visual stimuli and are important in the representation and recognition of visual images. Recording from 427 single neurons in the human hippocampus, entorhinal cortex and amygdala, we found a remarkable degree of category-specific firing of individual neurons on a trial-by-trial basis. Of the recorded neurons, 14% responded selectively to visual stimuli from different categories, including faces, natural scenes and houses, famous people and animals. Based on the firing rate of individual neurons, stimulus category could be predicted with a mean probability of error of 0.24. In the hippocampus, the proportion of neurons responding to spatial layouts was greater than to other categories. Our data provide direct support for the role of human medial temporal regions in the representation of different categories of visual stimuli.

Robustness and variability of neuronal coding by amplitude-sensitive afferents in the weakly electric fish eigenmannia

We investigated the variability of P-receptor afferent spike trains in the weakly electric fish, Eigenmannia, to repeated presentations of random electric field AMs (RAMs) and quantified its impact on the encoding of time-varying stimuli. A new measure of spike timing jitter was developed using the notion of spike train distances recently introduced by Victor and Purpura. This measure of variability is widely applicable to neuronal responses, irrespective of the type of stimuli used (deterministic vs. random) or the reliability of the recorded spike trains. In our data, the mean spike count and its variance measured in short time windows were poorly correlated with the reliability of P-receptor afferent spike trains, implying that such measures provide unreliable indices of trial-to-trial variability. P-receptor afferent spike trains were considerably less variable than those of Poisson model neurons. The average timing jitter of spikes lay within 1-2 cycles of the electric organ discharge (EOD). At low, but not at high firing rates, the timing jitter was dependent on the cutoff frequency of the stimulus and, to a lesser extent, on its contrast. When spikes were artificially manipulated to increase jitter, information conveyed by P-receptor afferents was degraded only for average jitters considerably larger than those observed experimentally. This suggests that the intrinsic variability of single spike trains lies outside of the range where it might degrade the information conveyed, yet still allows for improvement in coding by averaging across multiple afferent fibers. Our results were summarized in a phenomenological model of P-receptor afferents, incorporating both their linear transfer properties and the variability of their spike trains. This model complements an earlier one proposed by Nelson et al. for P-receptor afferents of Apteronotus. Because of their relatively high precision with respect to the EOD cycle frequency, P-receptor afferent spike trains possess the temporal resolution necessary to support coincidence detection operations at the next stage in the amplitude-coding pathway.

Tetanic stimulation leads to increased accumulation of Ca(2+)/calmodulin-dependent protein kinase II via dendritic protein synthesis in hippocampal neurons

mRNA for the alpha-subunit of CaMKII is abundant in dendrites of neurons in the forebrain (Steward, 1997). Here we show that tetanic stimulation of the Schaffer collateral pathway causes an increase in the concentration of alpha-CaMKII in the dendrites of postsynaptic neurons. The increase is blocked by anisomycin and is detected by both quantitative immunoblot and semiquantitative immunocytochemistry. The increase in dendritic alpha-CaMKII can be measured 100-200 micrometer away from the neuronal cell bodies as early as 5 min after a tetanus. Transport mechanisms for macromolecules from neuronal cell bodies are not fast enough to account for this rapid increase in distal portions of the dendrites. Therefore, we conclude that dendritic protein synthesis must produce a portion of the newly accumulated CaMKII. The increase in concentration of dendritic CaMKII after tetanus, together with the previously demonstrated increase in autophosphorylated CaMKII (Ouyang et al., 1997), will produce a prolonged increase in steady-state kinase activity in the dendrites, potentially influencing mechanisms of synaptic plasticity that are controlled through phosphorylation by CaMKII.

Characterization of the biosynthesis of beta(1-2) cyclic glucan in R. Fredii. Beta(1-2) glucan has no apparent role in nodule invasion of Mc Call and Peking soybean cultivars

Three wild type strains of Rhizobium fredii, USDA 191, USDA 257 and HH 303, do not synthesize in vivo or in vitro beta(1-3), beta(1-6) cyclic glucans, all strains form in vitro and in vivo cyclic beta(1-2) glucans. Approximately 80% of the recovered R. fredii cellular cyclic beta(1-2) glucans were anionic and the substituent was identified as phosphoglycerol. Inner membranes prepared from these R. fredii strains have a beta(1-2) glucan-intermediate-protein with apparent molecular mass undistinguishable from Agrobacterium tumefaciens beta(1-2) glucan intermediate protein. Studies of the degree of polymerization of the oligosaccharides recovered from the protein-intermediate after short pulse incubations with UDP-14C-glucose suggested that the rate limiting step in the biosynthesis of cyclic glucan is cyclization. Kinetic studies revealed that the K(m) for UDP-glucose was 0.33 mM. No difference was detected between the K(m) for initiation/elongation and cyclization reactions. Nodulation studies of a ndvB R. fredii mutant with Mc Call and Peking soybean cultivars, revealed that beta(1-2) glucans do not seem to be required for normal nodule invasion of these soybean cultivars.