Information encoding and the responses of single neurons in the primate temporal visual cortex.
M. J. TOVEE,E. T. ROLLS,A. TREVES,R. P. BELLIS
DOI: https://doi.org/10.1152/jn.1993.70.2.640
1993-01-01
Journal of Neurophysiology
Abstract:1. The possibility of temporal encoding in the spike trains of single neurons recorded in the temporal lobe visual cortical areas of rhesus macaques was analyzed with the use of principal compo- nent and information theory analyses of smoothed spike trains. The neurons analyzed had responses selective for faces. 2. Provided that a correction was applied to earlier methods of principal component analysis used for neuronal spike trains, it was shown that the first principal component provides by a great extent the most information, with the second and third adding only small proportions (on average 18.8 and 8.4%, respectively). 3. It was shown that the magnitude of the second and higher principal components is even smaller if the spike train analysis is started after the onset of the neuronal response, instead of before the neuronal response has started. This suggests that variations in response latency are at least a part of what is reflected by the second and higher principal components. 4. The first principal component was correlated with the mean firing rate of the neurons. The second and higher principal compo- nents reflected at least partly the onset properties of the neuronal responses, such as response latency differences between the stimuli. 5. A considerable proportion of the information available from principal components l-3 is available in the firing rate of the neuron. 6. Periods of the firing rate of as little as 50 or even 20 ms are sufficient to give a reasonable estimate of the firing rate of the neuron. 7. Information theory analysis showed that in short epochs (e.g., 50 ms) the information available from the firing rate can be as high, on average, as 84.4% of that available from the firing rate calculated over 400 ms, and 52.0% of that available from principal components l-3 in the 400-ms period. It was also found that 44.0% of the information calculated from the first three principal components is available in the firing rates calculated over epochs as short as 20 ms. 8. More information was available near the start of the neuro- nal response, and the information available from short epochs became less later in the neuronal response. 9. Taken together, these analyses provide evidence that a short period of firing taken close to the start of the neuronal response provides a reasonable proportion of the total information that would be available if a long period of neuronal firing (e.g., 400 ms) were utilized to extract it, even if temporal encoding were used. The implications of these and related findings are that, at least for rapid object recognition, each cortical stage provides information to the next in a short period of 20-50 ms, does not utilize temporal encoding, and completes sufficient computation to provide an output to the next stage in this same 20- to 50-ms period.