Neural data often span multiple indices, such as neuron, experimental condition, trial, and time, resulting in a tensor or multidimensional array. Standard approaches to neural data analysis often rely on matrix factorization techniques, such as principal component analysis or nonnegative matrix factorization. Any inherent tensor structure in the data is lost when flattened into a matrix. Here, we analyze datasets from primary motor cortex from the perspective of tensor analysis, and develop a theory for how tensor structure relates to certain computational properties of the underlying system. Applied to the motor cortex datasets, we reveal that neural activity is best described by condition-independent dynamics as opposed to condition-dependent relations to external movement variables. Motivated by this result, we pursue one further tensor-related analysis, and two further dynamical systems-related analyses. First, we show how tensor decompositions can be used to denoise neural signals. Second, we apply system identification to the cortex- to-muscle transformation to reveal the intermediate spinal dynamics. Third, we fit recurrent neural networks to muscle activations and show that the geometric properties observed in motor cortex are naturally recapitulated in the network model. Taken together, these results emphasize (on the data analysis side) the role of tensor structure in data and (on the theoretical side) the role of motor cortex as a dynamical system.

Prefrontal cortex (PFC) is critical to behavioral flexibility and, hence, the top-down control over bottom-up sensory information. The mechanisms underlying this capacity have been hypothesized to involve the propagation of alpha/beta (8-30 Hz) oscillations via feedback connections to sensory regions. In contrast, gamma (30-160 Hz) oscillations are thought to arise as a function of bottom-up, feedforward stimulation. To test the hypothesis that such oscillatory phenomena embody such functional roles, we assessed the performance of nine monkeys on tasks of learning, categorization, and working memory concurrent with recording of local field potentials (LFPs) from PFC. The first set of tasks consisted of two classes of learning: one, explicit and, another, implicit. Explicit learning is a conscious process that demands top-down control, and in these tasks alpha/beta oscillations tracked learning. In contrast, implicit learning is an unconscious process that is automatic (i.e. bottom up), and in this task alpha/beta oscillations did not track learning. We next looked at dot-pattern categorization. In this task, category exemplars were generated by jittering the dot locations of a prototype. By chance, some of these exemplars were similar to the prototype (low distortion), and others were not (high distortion). Behaviorally, the monkeys performed well on both distortion levels. However, alpha/beta band oscillations carried more category information at high distortions, while gamma-band category information was greatest on low distortions. Overall, the greater the need for top-down control (i.e. high distortion), the greater the beta, and the lesser the need (i.e. low distortion), the greater the gamma. Finally, laminar electrodes were used to record from animals trained on working memory tasks. Each laminar probe was lowered so that its set of contacts sampled all cortical layers. During these tasks, gamma oscillations peaked in superficial layers, while alpha/beta peaked in deep layers. Moreover, these deep-layer alpha/beta oscillations entrained superficial alpha/beta, and modulated the amplitude of superficial-layer gamma oscillations. These laminar distinctions are consistent with anatomy: feedback neurons originate in deep layers and feedforward neurons in superficial layers. In summary, alpha/beta oscillations reflect top-down control and feedback connectivity, while gamma oscillations reflect bottom-up processes and feedforward connectivity.

The physical architecture of neural circuits is thought to underlie the computations that give rise to higher order feature sensitivity in the neocortex. Recent technological breakthroughs have allowed the structural and functional investigation of the basic computational units of neural circuits; individual synaptic connections. However, it remains unclear how cortical neurons sample and integrate the thousands of synaptic inputs, supplied by different brain structures, to achieve feature selectivity. Here, I first describe how visual cortical circuits transform the elementary inputs supplied by the periphery into highly diverse, but well-organized, feature representations. By combining and optimizing newly developed techniques to map the functional synaptic connections with defined sources of inputs, I show that the intersection between columnar architecture and dendritic sampling strategies can lead to the selectivity properties of individual neurons: First, in the canonical feedforward circuit, the basal dendrites of a pyramidal neuron utilize unique strategies to sample ON (light increment) and OFF (light decrement) inputs in orientation columns to create the distinctive receptive field structure that is responsible for basic sensitivity to visual spatial location, orientation, spatial frequency, and phase. Second, for long-range horizontal connections, apical dendrites unbiasedly integrate functionally specialized and spatially targeted inputs in different orientation columns, which generates specific axial surround modulation of the receptive field. / Includes bibliography. / Dissertation (Ph.D.)--Florida Atlantic University, 2019. / FAU Electronic Theses and Dissertations Collection

Pavlovian fear conditioning describes a form of associative learning in which a previously neutral stimulus elicits a conditioned fear response after it has been temporally paired with an aversive consequence. Once acquired, the fear response can be extinguished by repeatedly presenting the former neutral stimulus in the absence of the aversive consequence. Although most patients suffering from anxiety disorders cannot recall a specific conditioned association between a formerly neutral stimulus and the feeling of anxiety, the produced behavioral symptoms, such as avoidance or safety behavior to prevent the anticipated aversive consequence are commonly exhibited in all anxiety disorders. Moreover, there is considerable similarity between the neural structures involved in fear and extinction in the rodent and in the human. Translational research thus contributes to the understanding of neural circuitries involved in the development and maintenance of anxiety disorders, and further provides hypotheses for improvements in treatment strategies aiming at inhibiting the fear response. Since the failure to appropriately inhibit or extinguish a fear response is a key feature of pathological anxiety, the present preclinical research focuses on the interplay between the amygdala and the medial prefrontal cortex (mPFC) during fear learning with particular regard to the prefrontal recruitment during fear extinction and its recall. By firstly demonstrating an increased mPFC activity over the time course of extinction learning with functional near-infrared spectroscopy, the main study of this dissertation focused on repetitive transcranial magnetic stimulation (rTMS) as brain stimulation technique suitable to enhance extinction learning. Since hypofrontality is assumed to underlie the maintenance of pathological anxiety, rTMS application revealed an increased mPFC activity, which resulted in a decreased fear response on the behavioral level both during extinction learning as well as during the recall of extinction 24 hours later and in the absence of another stimulation. The following attempt to improve the generalization of extinction with rTMS from an extinguished stimulus to a second stimulus which was reinforced but not extinguished was at least partially evidenced. By revealing an increased prefrontal activity to the non-extinguished stimulus, the active and the placebo rTMS condition, however, did not differ on behavioral parameters. These preclinical findings were discussed in the light of genetic and environmental risk factors with special regard to the combination of a risk variant of the neuropeptide S receptor 1 gene polymorphism (NPSR1 rs324981) and anxiety sensitivity. While the protective homozygous AA genotype group showed no correlation with anxiety sensitivity, the NPSR1 T genotype group exhibited an inverse correlation with anxiety sensitivity in the presence of emotionally negative stimuli. In light of other findings assuming a role of the NPSR1 T allele in panic disorder, the revealed hypofrontality was discussed to define a risk group of patients who might particularly benefit from an augmentation of exposure therapy with rTMS. Taken together, the presented studies support the central role of the prefrontal cortex in fear extinction and suggest the usefulness of rTMS as an augmentation strategy to exposure therapy in order to decrease therapy relapse rates. The combination of rTMS and extinction has been herein evidenced to modulate fear processes in a preclinical approach thereby establishing important implications for the design of future clinical studies. / Die Furchtkonditionierung nach Pavlov beschreibt einen assoziativen Lernmechanismus bei dem ein ursprünglich neutraler Stimulus nach wiederholter kontingenter Darbietung mit einem aversiven Stimulus zu einer konditionierten Furchtreaktion führt, die darauffolgend allein durch den nun konditionierten Reiz ausgelöst werden kann. Obwohl die meisten Angstpatienten keine initiale Reiz-Reaktionsverbindung erinnern können, gelten die Mechanismen der Furchtkonditionierung als Erklärungsmodelle für die Entstehung und Aufrechterhaltung von Angststörungen. Evidenz erhalten sie zudem durch den Einsatz und die Wirksamkeit expositionsbasierter Methoden in der Behandlung von Angststörungen. Ihnen liegt die Extinktion einer erworbenen konditionierten Reaktion zugrunde, bei der der konditionierte Reiz wiederholt ohne seine erwartete aversive Konsequenz dargeboten wird. Dies führt in der Folge zu einer abnehmenden Furchtreaktion. Da die neuronalen Strukturen, die in den Erwerb und die Extinktion einer konditionierten Furchtreaktion involviert sind, weitgehend speziesübergreifend sind, lassen sich aus Tiermodellen wertvolle Hypothesen zur Verbesserung bestehender Behandlungsstrategien mit dem Ziel der Reduktion der erworbenen Furchtreaktion generieren. Eine unzureichende Inhibition bzw. Extinktion der Furchtreaktion gilt als Charakteristikum von pathologischer Angst. Die im Rahmen dieser Dissertation vorgestellten Studien beschäftigen sich mit dem zugrundliegenden neurobiologischen Ungleichgewicht zwischen der Amygdala und dem Präfrontalkortex, das als ursächlich für die Aufrechterhaltung pathologischer Angst vermutet wird. Zunächst wird hierbei eine Untersuchung vorgestellt, bei der die zunehmende Beteiligung des Präfrontalkortex' über den Verlauf eines Extinktionstrainings erstmals mit der funktionellen Nahinfrarot- Spektroskopie dargestellt werden konnte. Da zunehmende Evidenz auf eine unzureichende präfrontale Kortexaktivierung bei pathologischer Angst hindeutet, beschäftigt sich die Hauptstudie dieser Dissertation mit der Fragestellung, ob die Aktivität des Präfrontalkortex' mit Hilfe der repetitiven transkraniellen Magnetstimulation (rTMS) gesteigert werden kann. In Analogie zu tierexperimentellen Untersuchungen konnte in einer Gruppe gesunder Probanden nach einer Stimulation mit rTMS verglichen mit einer Placebobedingung eine verringerte Furchtreaktion gezeigt werden, die auch während des Abrufs des Extinktionsgedächtnis nach 24 Stunden und unabhängig von einer erneuten Stimulation noch nachweisbar war. In einem nächsten Schritt wurde, wiederum in Anlehnung an tierexperimentelle Studien, die Generalisierung eines Extinktionstrainings auf einen ebenfalls konditionierten, aber nicht extingierten Stimulus untersucht. Hierbei zeigte sich eine partielle Bestätigung der Hypothesen. So konnten zwar auf behavioraler Ebene keine Gruppenunterschiede zwischen einer aktiven und einer Placebobedingung detektiert werden, in der aktiven Gruppe ließ sich 24 Stunden nach der Stimulation jedoch eine erhöhte präfrontale Kortexaktivierung auf den nicht-extingierten Stimulus zeigen. Diese Studienergebnisse werden auf Basis einer weiteren Arbeit zu Gen-Umwelt-Einflüssen diskutiert. Hierbei konnte eine Konstellation bestehend aus der Risikovariante (T Allel) des Neuropeptid S Rezeptor Gens (NPSR1 rs324981) und einer erhöhten Angstsensitivität im Unterschied zu einer homozygoten AA Genotyp-Gruppe mit einer verringerten präfrontalen Kortexaktivierung auf negative emotionale Stimuli assoziiert werden. Unter Einbezug des literarischen Kontexts zu NPSR1 und dem Auftreten der Panikstörung legen diese Ergebnisse nahe, dass insbesondere solche und ähnliche Risikogruppen von einer Augmentationsstrategie mit rTMS profitieren könnten. Zusammenfassend bestätigen die vorliegenden Studien die Rolle des Präfrontalkortex bei der Furchtextinktion und legen den Einsatz der rTMS für die Verbesserung der Expositionstherapie nahe. Aus diesen präklinischen Arbeiten werden Hinweise für die Umsetzung von klinischen Studien generiert, die über die Augmentation von Exposition mit rTMS zu einer Rückfallreduktion bei der Therapie von Angststörungen beitragen könnten.

Doctor of Philosophy / Binocular rivalry refers to the perceptual alternation that occurs while viewing incompatible images, in which one monocular image is dominant and the other is suppressed. Rivalry has been closely studied but the neural site at which it is initiated is still controversial. The central claim of this thesis is that primary visual cortex is responsible for its initiation. This claim is supported by evidence from four experimental studies. The first study (described in Chapter 4) introduces the methodology for measuring visual sensitivity during dominance and suppression and compares several methods to see which yields the greatest difference between these two sensitivities. Suppression depth was measured by comparing the discrimination thresholds to a brief test stimulus delivered during dominance and suppression phases. The deepest suppression was achieved after a learning period, with the test stimulus presented for 100 ms and with post-test masking. The second study (Chapter 5) compares two hypotheses for the mechanism of binocular rivalry. Under eye suppression, visibility decreases when the tested eye is being suppressed, regardless of the test stimulus’s features. Feature suppression, however, predicts that reduction of visibility is caused by suppression of a stimulus feature, no matter which eye is suppressed. Eye suppression claims that monocular channels in the visual system alternate between dominance and suppression, while Feature suppression assumes that the features of stimuli inhibit each other perceptually in the high-level cortex. The experiment used a test stimulus similar in features to one, but not the other, rivalry-inducing stimulus. Test sensitivity was found to be lowered when the test stimulus was presented to the eye whose rivalry-inducing stimulus was suppressed. Sensitivity was not lowered when the test stimulus was presented to the other eye, even when the test shared features with the suppressed stimulus. The conclusion is that feature suppression is weak or does not exist without eye suppression, and that rivalry therefore originates in the primary visual cortex. If binocular rivalry is initiated in the primary visual cortex, stimuli producing no coherent activity in that area should produce no rivalry. In the third study (Chapter 6) this idea was tested with rotating arrays of short-lifetime dots. The dots with the shortest lifetime produced an image with no rotation signal, and an infinite lifetime produced rigid rotation. Subjects could discriminate the rotation direction with high accuracy at all but the shortest lifetime. When the two eyes were presented with opposite directions of rotation, there was binocular rivalry only at the longest lifetimes. Stimuli with short lifetimes produce a coherent motion signal, since their direction can be discriminated, but do not produce rivalry. A simple interpretation of this observation is that binocular rivalry is initiated at a level in the visual hierarchy below that which supports the motion signal. The model supported by the results of previous chapters requires that binocular rivalry suppression be small in the primary visual cortex, and builds up as signals progress along the visual pathway. This model predicts that for judgements dependent on activity in high visual cortex: 1. Binocular rivalry suppression should be deep; 2. Responses should be contrast invariant. The fourth and last study (chapter 7) confirmed these predictions by measuring suppression depth in two ways. First, two similar forms were briefly presented to one eye: the difference in shapes required for their discrimination was substantially greater during suppression than during dominance. Second, the two forms were made sufficiently different in shape to allow easy discrimination at high contrast, and the contrast of these forms was lowered to find the discrimination threshold. The results in the second experiment showed that contrast sensitivity did not differ between the suppression and dominance states. This invariance in contrast sensitivity is interpreted in terms of steep contrast-response functions in cortex beyond the primary visual area. The work in this thesis supports the idea that binocular rivalry is a process distributed along the visual pathway. More importantly, the results provide several lines of evidence that binocular rivalry is initiated in primary visual cortex.

The capacity of the adult human nervous system to alter the strength of connections between neurons and between networks of neurons is an exciting area of research providing novel insights into the mechanisms involved in learning, memory and recovery following brain damage. In recent years, it has become clear that both afferent input into the motor cortex and the learning of a new motor task can drive cortical reorganisation. This thesis is concerned with the functional significance of this plasticity, in both normal subjects and stroke patients, and with the question of whether stimulation - induced plasticity can lead to improved fine motor performance. My initial experiments were conducted to determine the optimal method of analysing responses to transcranial magnetic stimulation ( TMS ), and to investigate aspects of motor performance as the hand performs a precision task to grasp and lift an object. Studies on normal subjects showed that there is little difference between the dominant and non - dominant hands performing this task, but the type of grip used influences grip - force control. An investigation of stroke patients performing this task demonstrated that certain parameters were sensitive to differences between the affected and unaffected hands and these parameters were highly correlated with stroke - specific functional outcome measures. The induction of plastic change in the human motor cortex can be induced by repetition of movements, performing a complex motor task or stimulation of the peripheral afferents and / or the motor cortex itself. I observed that the application of so - called " associative stimulation " to two hand muscles in normal subjects increased the excitability of the corticospinal projection to those muscles, and improved performance times on a subsequent motor task to a greater extent than subjects receiving a control intervention. I then applied associative stimulation to the affected hand of stroke patients in conjunction with rehabilitation, which improved their ability to perform the dextrous grip - lift task. This is the first study to show that this method of inducing motor cortical plasticity can also lead to functional improvements in stroke patients. These studies confirm that using afferent stimulation to drive cortical reorganisation is associated with improved function and fine motor performance in both normal subjects and stroke patients. / Thesis (Ph.D.)--School of Molecular and Biomedical Science, 2006.

N/A

"April 2003" / Bibliography: leaves 124-144. / xvii, 144 leaves : ill. ; 30 cm. / Title page, contents and abstract only. The complete thesis in print form is available from the University Library. / Thesis (Ph.D.)--University of Adelaide, School of Molecular and Biomedical Sciences, Discipline of Physiology, 2003

Thesis (Ph. D.)--Ohio State University, 2007. / Title from first page of PDF file. Includes bibliographical references (p. 131-154).

This paper sketches a hypothetical cortical architecture for visual 3D object recognition based on a recent computational model. The view-centered scheme relies on modules for learning from examples, such as Hyperbf-like networks. Such models capture a class of explanations we call Memory-Based Models (MBM) that contains sparse population coding, memory-based recognition, and codebooks of prototypes. Unlike the sigmoidal units of some artificial neural networks, the units of MBMs are consistent with the description of cortical neurons. We describe how an example of MBM may be realized in terms of cortical circuitry and biophysical mechanisms, consistent with psychophysical and physiological data.