Christoph Rasche

Silicon synaptic depression.

Abstract.u2002The recent quantitative description of activity-dependent depression in the synaptic transmission between cortical neurons has lead to many interesting suggestions of possible computational implications. Based on a simple biological model, we have constructed an analog circuit that emulates the properties of short-term depressing synapses. The circuit comprises only seven transistors and two capacitors per synapse, and is able to reproduce computational features of depressing synapses such as the 1/F law, the detection of long intervals of presynaptic silence and the sensitivity to redistribution of presynaptic firing rates. It provides a useful basis for implementing neural networks with dynamical synapses.

Precision of speed discrimination and smooth pursuit eye movements

Several studies have shown that the precision of smooth pursuit eye speed can match perceptual speed discrimination thresholds during the steady-state phase of pursuit [Kowler, E., & McKee, S. (1987). Sensitivity of smooth eye movement to small differences in target velocity. Vision Research, 27, 993-1015; Gegenfurtner, K., Xing, D., Scott, B., & Hawken, M. (2003). A comparison of pursuit eye movement and perceptual performance in speed discrimination. Journal of Vision, 3, 865-876]. Recently, Osborne et al. [Osborne, L. C., Lisberger, S. G., & Bialek, W. (2005). A sensory source for motor variation. Nature, 437, 412-416; Osborne, L. C., Hohl, S. S., Bialek, W., & Lisberger S. G. (2007). Time course of precision in smooth-pursuit eye movements of monkeys. Journal of Neuroscience, 27, 2987-2998] claimed that pursuit precision during the initiation phase of pursuit also matches the sensory variability, implying that there is no motor noise added during pursuit initiation. However, these results were derived from a comparison of monkey pursuit data to human perceptual data from the literature, which were obtained with different stimuli. To directly compare precision for perception and pursuit, we measured pursuit and perceptual variability in the same human observers using the same stimuli. Subjects had to pursue a Gaussian blob in a step-ramp paradigm and give speed judgments on the same or in different trials. Speed discrimination thresholds were determined for different presentation durations. The analysis of pursuit precision was performed for short intervals containing the initiation period only and also for longer intervals including steady-state pursuit. In agreement with published studies, we found that the Weber fractions for psychophysical speed discrimination were fairly constant for different presentation durations, even for the shortest presentation duration of 150ms. Pursuit variability was 3-4 times as high for the analysis interval (300ms) containing the open-loop phase only. For pursuit analysis intervals of 400-500ms, pursuit variability approached perceptual variability. Our results show that, for the stimuli we used, the motor system contributes at least 50% to the total variability of smooth pursuit eye movements during the initiation phase.

Smooth Pursuit Eye Movements to Isoluminant Targets

At slow speeds, chromatic isoluminant stimuli are perceived to move much slower than comparable luminance stimuli. We investigated whether smooth pursuit eye movements to isoluminant stimuli show an analogous slowing. Beside pursuit speed and latency, we studied speed judgments to the same stimuli during fixation and pursuit. Stimuli were either large sine wave gratings or small Gaussians blobs moving horizontally at speeds between 1 and 11 degrees /s. Targets were defined by luminance contrast or color. Confirming prior studies, we found that speed judgments of isoluminant stimuli during fixation showed a substantial slowing when compared with luminance stimuli. A similarly strong and significant effect of isoluminance was found for pursuit initiation: compared with luminance targets of matched contrasts, latencies of pursuit initiation were delayed by 50 ms at all speeds and eye accelerations were reduced for isoluminant targets. A small difference was found between steady-state eye velocities of luminance and isoluminant targets. For comparison, we measured latencies of saccades to luminance and isoluminant stimuli under similar conditions, but the effect of isoluminance was only found for pursuit. Parallel psychophysical experiments revealed that different from speed judgments of moving isoluminant stimuli made during fixation, judgments during pursuit are veridical for the same stimuli at all speeds. Therefore information about target speed seems to be available for pursuit eye movements and speed judgments during pursuit but is degraded for perceptual speed judgments during fixation and for pursuit initiation.

Forward- and backpropagation in a silicon dendrite

We have developed an analog very-large-scale integrated (aVLSI) electronic circuit that emulates a compartmental model of a neuronal dendrite. The horizontal conductances of the compartmental model are implemented as a switched capacitor network. The transmembrane conductances are implemented as transconductance amplifiers. The electrotonic properties of our silicon cable are qualitatively similar to those of the ideal passive cable that is commonly used to model mathematically the electrotonic behavior of neurons. In particular the propagation of excitatory postsynaptic potentials is realistic, and we are easily able to emulate such classical synaptic integration models as direction selectivity. We are also able to emulate the backpropagation into the dendrite of single somatic spikes and bursts of spikes. Thus, this silicon dendrite is suitable for incorporation in detailed silicon neurons operating in real-time; in particular for the emulation of forward- and backpropagating electrical activities found in real neurons.

An Improved Silicon Neuron

We describe an improved spiking silicon neuron (SN) [6] that approximates the dynamics of ionic currents of a real nerve cell. The improved version has less circuitry and fewer parameters than previous circuits thereby improving the spiking characteristics. We describe the differential equations governing the revised circuits and use them to explain the spiking properties of the SN. We also describe how to tune the parameters efficiently to bring the neuron quickly into its correct operating regime. The new neurons are sufficiently robust for operation in large networks. We demonstrate their robustness by comparing the neurons frequency-current curve between different chips for the same set of parameter values.

An Approach to the Parameterization of Structure for Fast Categorization

A decomposition is described, which parameterizes the geometry and appearance of contours and regions of gray-scale images with the goal of fast categorization. To express the contour geometry, a contour is transformed into a local/global space, from which parameters are derived classifying its global geometry (arc, inflexion or alternating) and describing its local aspects (degree of curvature, edginess, symmetry). Regions are parameterized based on their symmetric axes, which are evolved with a wave-propagation process enabling to generate the distance map for fragmented contour images. The methodology is evaluated on three image sets, the Caltech 101 set and two sets drawn from the Corel collection. The performance nearly reaches the one of other categorization systems for unsupervised learning.

Recognizing the gist of a visual scene: possible perceptual and neural mechanisms

We try to understand the basics of human image processing from a gist recognition perspective. Because the gist is only a subset of the images information, we think that it is extracted with help of interpretation (feedback). In a perceptual section we list possible mechanisms that the interpretation process uses to determine the gist: in addition to the commonly known local-to-global perception evolvement, there is likely to be also a global-to-local evolvement direction, a coarse/fine scale, as well as a foreground/background scale. In a neural section we first summarize feedback connections that can possibly be involved in gist recognition. Second, we propose that the perceptual mechanisms are spread all over the cortex and that cortical visual computation occurs distributively rather than hierarchical.

Content-based video description for automatic video genre categorization

In this paper, we propose an audio-visual approach to video genre categorization. Audio information is extracted at block-level, which has the advantage of capturing local temporal information. At temporal structural level, we asses action contents with respect to human perception. Further, color perception is quantified with statistics of color distribution, elementary hues, color properties and relationship of color. The last category of descriptors determines statistics of contour geometry. An extensive evaluation of this multi-modal approach based on on more than 91 hours of video footage is presented. We obtain average precision and recall ratios within [87%−100%] and [77%−100%], respectively, while average correct classification is up to 97%. Additionally, movies displayed according to feature-based coordinates in a virtual 3D browsing environment tend to regroup with respect to genre, which has potential application with real content-based browsing systems.

Silicon Synaptic Conductances

We have developed compact analog integrated circuits that simulate two synaptic excitatory conductances. A four-transistor circuit captures the dynamics of an excitatory postsynaptic current caused by a real AMPA conductance. A six-transistor circuit simulates the effects of a real voltage-dependent NMDA conductance. The postsynaptic current dynamics are modeled by a current mirror integrator with adjustable gain. The voltage dependence of the silicon NMDA conductance is realized by a differential pair. We show the operation of these silicon synaptic conductances and their integration with the silicon neuron (Mahowald and Douglas, 1991).

Neuromorphic Excitable Maps for Visual Processing

An excitable membrane is described which can perform different visual tasks such as contour detection, contour propagation, image segmentation, and motion detection. The membrane is designed to fit into a neuromorphic multichip system. It consists of a single two-dimensional (2-D) layer of locally connected integrate-and-fire neurons and propagates input in the subthreshold and the above-threshold range. It requires adjustment of only one parameter to switch between the visual tasks. The performance of two spiking membranes of different connectivity is compared, a hexagonally and an octagonally connected membrane. Their hardware and system suitability is discussed