Cortical mapping of gamma oscillations in areas V1 and V4 of the macaque monkey GABRIEL ROLS, 1 CATHERINE TALLON-BAUDRY, 2 PASCAL GIRARD, 1 OLIVIER BERTRAND, 2 and JEAN BULLIER 1 1 INSERM u371 69500 Bron, France 2 INSERM u280, Mental Processes and Brain Activity, Lyon, France (Received November 13, 2000; Accepted April 2, 2001) Abstract To characterize the temporal and spatial parameters of gamma activity evoked by visual stimuli in areas V1 and V4 of the monkey cortex, we recorded the electrocorticogram (ECoG) with an implanted array of 28 and 31 subdural electrodes placed over the surface of the operculum in two anesthetized monkeys. This intermediate level of recordings should help to bridge the gap between multiunit and scalp recordings. Both averaged and single-trial responses to small flashed stimuli, for which we varied the retinotopic position, the luminance and the color, were analyzed in the time-frequency domain using a wavelet-based decomposition of the signal. Large gamma oscillations (40–55 Hz), not phase locked to stimulus onset, were observed during the whole stimulus presentation, whereas visual evoked potentials (VEPs) were present mainly at stimulus onset and offset. Cortical mapping showed that both activities were restricted in spatial extent and followed the retinotopic organization of area V1 on the operculum, thus strongly suggesting they were generated in the underlying cortex. Oscillatory burst detection in single trials showed that one to two bursts lasting from 100 ms to 500 ms occurred in the first 500 ms following stimulus onset, and that bursts occurring during the subsequent phases of the response had a smaller amplitude and duration. Finally, we showed that gamma activity was stronger with higher luminances and for red than for green, yellow, or white stimuli of same luminance. In one animal we recorded gamma activity over area V4. This was of lower magnitude than the activity recorded over V1 and was delayed by 40 ms with respect to the beginning of gamma activity in V1, in contrast with the VEPs that were delayed by 20 ms only. Both gamma oscillations and early VEP followed the retinotopic organization of V4 over the prelunate gyrus. The results show that gamma oscillations are dependent upon the same parameters as the VEPs (retinotopic position, luminance, and color). However, the differences in the time course of VEPs and gamma activity (transient vs. sustained) suggests that these two responses may reflect different cell populations, different networks, or different firing modes. Keywords: Electrocorticogram, Mapping, Vision, Oscillations, Synchronization Introduction In the last decade, much attention has been paid to oscillatory synchronization in the visual system, as a possible solution to the binding problem (Gray, 1999; Singer, 1999): neurons responding to different parts of the same object would synchronize their discharges on an oscillatory mode in the gamma range (20–100 Hz). Oscillatory synchronization is more likely to be best observed in population signals such as local field potentials (LFPs), since this type of signal reflects the coordinated activity of a small neural ensemble. Indeed, long before any functional relevance of tempo- ral coding was put forward, prominent oscillations were reported in response to a flash from the visual cortex of humans (Chatrian et al., 1960) and alert monkeys (Hughes, 1964). However, many of the recent studies on synchronization in the visual cortex of cats and monkeys relied solely on the quantification of multiunit ac- tivity (MUA) (Kreiter & Singer, 1996; Kruse & Eckhorn, 1996; Livingstone, 1996; Castelo-Branco et al., 1998; Herculano-Houzel et al., 1999), even though a close correspondence between epi- sodes of oscillatory activity in the LFPs and in the multiunit activity recorded from the same electrode is sometimes mentioned (Gray & Singer, 1989; Engel et al., 1990; Kruse & Eckhorn, 1996; Livingstone, 1996). Gamma oscillations in intracortical LFPs in response to a visual stimulus have already been observed and analyzed (Freeman & Van Dijk, 1987; Juergens et al., 1999); however, much remains to be learned of their detailed time and frequency contents, as well as of their spatial organization. In Address correspondence and reprint requests to: Catherine Tallon- Baudry, INSERM u280, 151, cours Albert Thomas, F-69003 Lyon, France. E-mail: tallon-baudry@lyon151.inserm.fr Present address of G. Rols, P. Girard and J. Bullier: Centre de Recher- che Cerveau et Cognition, CNRS & Université Paul Sabatier, Toulouse, France. Visual Neuroscience (2001), 18, 527–540. Printed in the USA. Copyright © 2001 Cambridge University Press 0952-5238001 $12.50 DOI: 10.1017.S0952523801184038 527