Active Filtering: A Predictive Function of Recurrent Circuits of Sensory Cortex.

Our brains encode many features of the sensory world into memories: we can sing along with songs we have heard before, interpret spoken and written language composed of words we have learned, and recognize faces and objects. Where are these memories stored? Each area of the cerebral cortex has a huge number of local, recurrent, excitatory–excitatory synapses, as many as 500 million per cubic millimeter. Here I review evidence that cortical recurrent connectivity in sensory cortex is a substrate for sensory memories. Evidence suggests that the local recurrent network encodes the structure of natural sensory input, and that it does so via active filtering, transforming network inputs to boost or select those associated with natural sensation. This is a form of predictive processing, in which the cortical recurrent network selectively amplifies some input patterns and attenuates others, and a form of memory.

BackgroundThe organization of the connectivity between mammalian cortical areas has become a major subject of study, because of its important role in scaffolding the macroscopic aspects of animal behavior and intelligence. In this study we present a computational reconstruction approach to the problem of network organization, by considering the topological and spatial features of each area in the primate cerebral cortex as subsidy for the reconstruction of the global cortical network connectivity. Starting with all areas being disconnected, pairs of areas with similar sets of features are linked together, in an attempt to recover the original network structure.ResultsInferring primate cortical connectivity from the properties of the nodes, remarkably good reconstructions of the global network organization could be obtained, with the topological features allowing slightly superior accuracy to the spatial ones. Analogous reconstruction attempts for the C. elegans neuronal network resulted in substantially poorer recovery, indicating that cortical area interconnections are relatively stronger related to the considered topological and spatial properties than neuronal projections in the nematode.ConclusionThe close relationship between area-based features and global connectivity may hint on developmental rules and constraints for cortical networks. Particularly, differences between the predictions from topological and spatial properties, together with the poorer recovery resulting from spatial properties, indicate that the organization of cortical networks is not entirely determined by spatial constraints.

mGluR5-dependent increases in immediate early gene expression in the rat striatum following acute administration of amphetamine

Abstract— This study was carried out to ascertain what biochemical changes might be present in cobalt‐induced epilepsy in the rat. Sodium, potassium, calcium, magnesium, Na‐K ATPase activity, water content, protein content and the ability of the tissue to utilize oxygen were measured in (1) the area of the cerebral cortex in which the cobalt was implanted; (2) in an area adjacent to but not including the area of the lesion; and (3) in the homotopic area of the contralateral cerebral cortex. The greatest changes were observed in the area of the lesion itself, with marked increases in calcium, magnesium and sodium contents and decreases in potassium content, Na‐K ATPase activity, protein content and the ability of the tissue to utilize oxygen. The only significant findings in the area adjacent to the lesion and in the contralateral cortex were a modest elevation of sodium and a modest decrease in potassium at different time periods after implantation of the cobalt. We feel that the changes observed at the site of cobalt implantation may reflect tissue destruction which is unrelated to the epileptic process.

Functional magnetic resonance imaging (fMRI) was used to investigate activation of the multimodal areas in the cerebral cortex–supramarginal and angular gyri, precuneus, and middle temporal visual cortex (MT/V5)–in response to motion of biologically significant sounds (human footsteps). The subjects listened to approaching or receding footstep sounds during 45 s, and such stimulation was supposed to evoke auditory adaptation to biological motion. Listening conditions alternated with stimulation-free control. To reveal activity in the regions of interest, the periods before and during stimulation were compared. Most stable and voluminous activation was detected in the supramarginal and angular gyri, being registered for all footstep sound types–approaching, receding and steps in place. Listening to human approaching steps activated the precuneus area, with the volume of activation clusters varying considerably between subjects. In the MT/V5 area, activation was revealed in 5 of 21 subjects. The involvement of the tested multimodal cortical areas in analyzing biological motion is discussed.

The organization of multiple motor areas in the cerebral cortex has been investigated frequently in primates but rarely in nonprimate species. To compare sensorimotor areas in cats and primates, the cytoarchitecture of frontal and parietal areas of the cat cerebral cortex was described and mapped from coronal sections stained with cresyl violet or thionine. Multiple subdivisions of areas 4 and 6 were recognized; of these, the cytoarchitecture of area 4 gamma is similar to that of area 4 described in other carnivores and in primates and is characterized by giant pyramidal cells in multiple rows or clusters in lamina V. In other subdivisions of area 4 (4 delta, 4sfu, and 4fu), giant pyramidal cells are few or absent in lamina V, and these subdivisions resemble area 6 of primates. Area 6 of the cat cortex is heterogeneous, and differences in laminar appearance and size of pyramidal cells in lamina V distinguish its four subdivisions (6a alpha, 6a beta, 6a gamma, and 6iffu). The adjoining prefrontal areas are distinguishable from area 6 by the presence of a thin internal granular lamina (lamina IV) and the reduced size of pyramidal cells in lamina V. Laminae are poorly differentiated in the cingulate areas, where a rostral and caudal subdivision can be distinguished on the basis of the absence or presence of lamina IV. Area 3a is characterized by a thin lamina IV and is located between frontal agranular and parietal granular (well-defined lamina IV) fields (3b, 1, 2, 2pri, 5, and 7). Insular cortex can be subdivided into granular and agranular fields.

Modern medicine has generally viewed the concept of "psychosomatic" disease with suspicion. This view arose partly because no neural networks were known for the mind, conceptually associated with the cerebral cortex, to influence autonomic and endocrine systems that control internal organs. Here, we used transneuronal transport of rabies virus to identify the areas of the primate cerebral cortex that communicate through multisynaptic connections with a major sympathetic effector, the adrenal medulla. We demonstrate that two broad networks in the cerebral cortex have access to the adrenal medulla. The larger network includes all of the cortical motor areas in the frontal lobe and portions of somatosensory cortex. A major component of this network originates from the supplementary motor area and the cingulate motor areas on the medial wall of the hemisphere. These cortical areas are involved in all aspects of skeletomotor control from response selection to motor preparation and movement execution. The second, smaller network originates in regions of medial prefrontal cortex, including a major contribution from pregenual and subgenual regions of anterior cingulate cortex. These cortical areas are involved in higher-order aspects of cognition and affect. These results indicate that specific multisynaptic circuits exist to link movement, cognition, and affect to the function of the adrenal medulla. This circuitry may mediate the effects of internal states like chronic stress and depression on organ function and, thus, provide a concrete neural substrate for some psychosomatic illness.

Rapid Neocortical Dynamics: Cellular and Network Mechanisms

Recurrent neuronal circuits in the neocortex

This paper discusses the application of a shunt active power filter with energy storage, for voltage and power improvements within a local power supply network (micro-grid) in presence of distributed generation. Theoretical analysis investigates the control the shunt active power filter in response to voltage sag. The shunt active filter operates over a short period of time, and is coordinated with the response of a distributed generator with a slower dynamic response. Simulation results show that the proposed method makes it possible to regulate the grid voltage at its nominal value and also allows the power sharing between the active filter and the local synchronous generator

Inhibitory effect of acupuncture on neuronal apoptosis in rats after cerebral ischemia

Neuronal Selectivity and Local Map Structure in Visual Cortex

A basic feature of cortical networks is a large degree of divergence and convergence, both locally and between cortical areas. Each cortical pyramidal cell receives approximately 10,000 synaptic inputs, of which about 75% are excitatory 1–3. The vast majority of these excitatory inputs arise from other cortical neurons, with each presynaptic cell donating only a few synapses, leading to a very large degree of convergence of synaptic input onto each cortical cell. Since cortical pyramidal cells typically give rise to a dense innervation of the local region in which they reside, a high degree of local, recurrent connectivity is generated. Pyramidal to pyramidal cell recurrent connectivity is expressed within a cortical lamina as well as between cortical laminae, resulting in a local horizontal and vertical interconnected network. Interposed between this recurrent excitatory network of pyramidal cells are a large variety of GABAergic inhibitory interneurons4,5. One prominent pattern of connectivity in the cerebral cortex is that local axons from pyramidal cells innervate not only other pyramidal cells, but also GABAergic interneurons, which in turn innervate local pyramidal cells. This pattern of connectivity places GABAergic neurons in the role of regulator of excitatory communication within the cortex. Indeed, reduction or loss of this inhibitory control quickly results in the generation of the abnormal and powerful synchronized cortical discharges of epilepsy6.

Motor maps and electrical thresholds for evoking movements from motor areas of the cerebral cortex were evaluated in normal cats by using intracortical microstimulation techniques. Stainless steel chambers were implanted over craniotomies in adult cats trained to perform reaching and retrieval movements with their forelimbs. Prehensile motor training was continued and movement performance monitored for about 6–10 weeks during which the cortex was progressively explored with sharp tungsten electrodes inserted into cortical gyri (anterior and posterior sigmoid, and coronal) and the banks of sulci (cruciate, presylvian and coronal). Twice weekly, under light general anaesthesia, 3–4 tracks were made in either hemisphere till about 50 tracks were made in each hemisphere. Mean thresholds for evoking forelimb movements from different cytoarchitectonic areas (4γ, 4δ, 6aγ and 3a) were compared and no consistent or significant differences were observed between the different areas. In the animals (4/6) which used either forelimb to perform the tasks, there were no consistent differences in the mean thresholds for evoking forelimb movements from the two hemispheres. However, in 2 animals, which used their right forelimbs predominantly or exclusively to perform all the tasks, mean thresholds for evoking forelimb movements was significantly higher in areas 4γ and 6aγ of the left hemisphere (compared to the right); no consistent differences in the mean thresholds for evoking hindlimb or facial movements were observed between the two hemispheres. These findings suggest that ICMS thresholds for evoking forelimb movements may be similar in different sensorimotor areas of the cat cerebral cortex, and these thresholds could be influenced by motor training.

Peripheral and Central Inputs Shape Network Dynamics in the Developing Visual Cortex In Vivo