NEURAL ENSEMBLE CODING IN INFERIOR TEMPORAL CORTEX

1. Isolated, single-neuron extracellular potentials were recorded sequ entially in area TE of the inferior temporal cortex (IT) of two macaqu e monkeys (n = 58 and n = 41 neurons). Data were obtained while the an imals were performing a paired-associate task. The task utilized five stimuli and eight stimulus pairings (4 correct and 4 incorrect). Data were evaluated as average spike rate during experimental epochs of 100 or 400 ms. Single-unit and population characteristics were measured u sing a form of linear discriminant analysis and information theoretic measures. To evaluate the significance of covariance on population cod e measures, additional data consisting of simultaneous recordings from less than or equal to 8 isolated neurons(n = 37) were obtained from a third macaque monkey that was passively viewing visual stimuli. 2. On average, 43% of IT neurons were activated by any of the stimuli used (60% if those inhibited also are included). Yet the neurons were rathe r unique in the relative magnitude of their responses to each stimulus in the test set. These results suggest that information may be repres ented in IT by the pattern of activity across neurons and that the rep resentation is not sparsely coded. It is further suggested that the re presentation scheme may have similarities to DNA or computer codes whe rein a coding element is not a local parametric descriptor. This is a departure from the V1 representation, which appears to be both local a nd parametric. It is also different from theories of IT representation that suggest a constructive basis set or ''alphabet''. From this view , determination of stimurus discrimination capacity in IT should be ev aluated by measures of population activity patterns. 3. Evaluation of small groups of simultaneously recorded neurons obtained during a fixa tion task suggests that little information about visual stimuli is con veyed by covariance of activity in IT when a 100-ms time scale is used as in this study. This finding is consistent with a prior report, by Gochin et al., which used a I-ms time scare and failed to find neural activity coherence or oscillations dependent on stimuli. 4. Population -stimulus-discrimination capacity measures were influenced by the numb er of neurons and to some extent the number and type of stimuli. 5. In formation conveyed by individual neurons (mutual information) averaged 0.26 bits. The distribution of information values was uni-modal and i s therefore more consistent with a distributed than a local coding sch eme. Populations of neurons conveyed less information than the sum of the individuals, reaching <1.9 bits for similar to 50 neurons, thus in dicating considerable redundancy. The novelty of information was found to be a function of the size of the population sample, apparently dec reasing by 1/root n. 6. Information available through the course of th e behavioral paradigm was evaluated by partitioning the data into 100- ms epochs. The results indicate that a relatively large amount of info rmation becomes available in the second 100 ms after stimulus onset. T he peak in information occurs between 200 and 400 ms after stimulus on set and significantly decays before the stimulus is removed. By the ti me the second stimulus was shown in our paired associates paradigm (50 0 ms after stimulus 1 offset), information about the first stimulus wa s about half of the peak value.