Biomedical subjects

Darragh Smyth

Publications and source records attributed to Darragh Smyth.

4 recordsLinked to original sources

PubMed2003

The receptive-field organization of simple cells in primary visual cortex of ferrets under natural scene stimulation.

The responses of simple cells in primary visual cortex to sinusoidal gratings can primarily be predicted from their spatial receptive fields, as mapped using spots or bars. Although this quasilinearity is well documented, it is not clear whether it holds for complex natural stimuli. We recorded from simple cells in the primary visual cortex of anesthetized ferrets while stimulating with flashed digitized photographs of natural scenes. We applied standard reverse-correlation methods to quantify the average natural stimulus that invokes a neuronal response. Although these maps cannot be the receptive fields, we find that they still predict the preferred orientation of grating for each cell very well (r = 0.91); they do not predict the spatial-frequency tuning. Using a novel application of the linear reconstruction method called regularized pseudoinverse, we were able to recover high-resolution receptive-field maps from the responses to a relatively small number of natural scenes. These receptive-field maps not only predict the optimum orientation of each cell (r = 0.96) but also the spatial-frequency optimum (r = 0.89); the maps also predict the tuning bandwidths of many cells. Therefore, our first conclusion is that the tuning preferences of the cells are primarily linear and constant across stimulus type. However, when we used these maps to predict the actual responses of the cells to natural scenes, we did find evidence of expansive output nonlinearity and nonlinear influences from outside the classical receptive fields, orientation tuning, and spatial-frequency tuning.

Action Potentials↗

PubMed2003

Methods for first-order kernel estimation: simple-cell receptive fields from responses to natural scenes.

Recent studies have recovered receptive-field maps of simple cells in visual cortex from their responses to natural scene stimuli. Natural scenes have many theoretical and practical advantages over traditional, artificial stimuli; however, the receptive-field estimation methods are more complex than for white-noise stimuli. Here, we describe and justify several of these methods-spectral correction of the reverse correlation estimate, direct least-squares solution, iterative least-squares algorithms and regularized least-squares solutions. We investigate the pros and cons of the different methods, and evaluate them in a head-to-head comparison for simulated simple-cell data. This shows that, at least for quasilinear simulated simple cells, a regularized solution ('reginv') is most efficient, requiring fewer stimulus presentations for high-resolution reconstruction of the first-order kernel. We also investigate several practical issues that determine the success of this kind of experiment-the effects of neuronal nonlinearities, response variability and the choice of stimulus regime.

Nature↗

PubMed2002

Visual experience before eye-opening and the development of the retinogeniculate pathway.

Visual experience before eye-opening is not usually thought to have any developmental significance. Here we show that naturalistic visual stimuli presented through unopened eyelids robustly activate neurons in the ferret dorsal lateral geniculate nucleus. Further, dark-rearing prior to natural eye-opening has striking effects upon geniculate physiology. Receptive field maps after dark-rearing show increased convergence of On- and Off-center responses, and neurons frequently respond to both bright and dark phases of drifting gratings. There is also increased selectivity for the orientation of the gratings. These abnormalities of On-Off segregation can be explained by the finding that the responses of immature On and Off cells to naturalistic stimuli are strongly anticorrelated.

Action Potentials↗

PubMed1998

Local online learning of coherent information.

One of the goals of perception is to learn to respond to coherence across space, time and modality. Here we present an abstract framework for the local online unsupervised learning of this coherent information using multi-stream neural networks. The processing units distinguish between feedforward inputs projected from the environment and the lateral, contextual inputs projected from the processing units of other streams. The contextual inputs are used to guide learning towards coherent cross-stream structure. The goal of all the learning algorithms described is to maximize the predictability between each unit output and its context. Many local cost functions may be applied: e.g. mutual information, relative entropy, squared error and covariance. Theoretical and simulation results indicate that, of these, the covariance rule (1) is the only rule that specifically links and learns only those streams with coherent information, (2) can be robustly approximated by a Hebbian rule, (3) is stable with input noise, no pairwise input correlations, and in the discovery of locally less informative components that are coherent globally. In accordance with the parallel nature of the biological substrate, we also show that all the rules scale up with the number of streams.

Journal Article↗