The functional nature of neuronal oscillations

neuronal oscillations Singer and his colleagues have reported that oscillations of 30 70Hz occur in cat and monkey visual cortex 1'2 that can be synchronous across cortical columns 3. They have suggested that these oscillations may form the substrate of a temporal binding code by becoming synchronized in different populations of cells that process different aspects of the same stimulus 4. However, there are a number of points that should be considered. First, it has been suggested that their methods of statistical analysis may lead to an overestimation of the number of cells showing oscillatory activity 5. The implication is that the number of cells exhibiting oscillations in the cat visual cortex is probably much lower than Singer and his colleagues suggest. In addition, reliable oscillations are even more difficult to demonstrate in the monkey visual cortex 2'5'6. Second, the frequency and amplitude of oscillations seem to decrease as the velocity of a stimulus decreases 7. Indeed, we have found no evidence for oscillations in the monkey temporal visual cortex using static stimuli 6. These findings suggest that oscillations are not involved in temporal binding in primates when information on static objects is processed. Third, temporal constraints may also minimize the usefulness of oscillations for the processing of visual stimuli. There may only be lO-20ms of processing time at each visual cortical area for tasks such as object recognition 6'8'9. Given that even a strongly stimulated visual cell fires at only about 100 Hz, there will be time for only one or two spikes from each active neuron to be used in information processing in each cortical area that is required for recognition 6'1°. This suggests that oscillations that develop over tens of milliseconds and persist for hundreds of milliseconds in the visual system are unlikely to be crucial for object recognition. In conclusion, it can be argued that the low proportion of cells displaying the oscillation phenomenon, its absence in the monkey in response to static visual stimuli, and the temporal constraints on the processing of visual information suggest that the oscillations do not form a basis for spatial binding, at least in the processing of static stimuli, which might be achieved in other ways 11. Martin J. Tov~e Edmund T. Rolls Dept of Experimental Psychology, University of Oxford, South Parks Road, Oxford, UK OXl 3UD.