2. Action potentials from single cells were recorded extracellularly and isolated on the basis of amplitude. Receptive-field properties of the neurons under study were characterized initially by their discharge in response to gratings of sinusoidal luminance. By varying orientation and spatial frequency, optimal stimulus characteristics were determined. Oscillation analysis was performed on spike trains acquired during repeated presentations of the optimal stimulus by identification of power spectra peaks in the frequency range of rhythmic potentials observed in electroencephalograph studies (30-80 Hz). The amplitude and frequency of the largest peak in this range was used to characterize oscillatory strength and frequency. All discharge in which the peak amplitude exceeded the high-frequency noise by a factor > 1.5 was classified as oscillatory.
3. Of the 342 cortical cells examined, 147 cells displayed oscillatory activity in the 30 to 80-Hz range during portions of their visual response. Sixty out of 169 simple cells, 82 out of 166 complex cells, and 5 out of 7 special complex cells exhibited oscillations. There was no laminar bias in the distribution of oscillatory cells; the proportions of oscillatory cells were similar in all layers. All oscillatory discharge was variable with respect to frequency and strength between successive presentations of the same optimal stimulus. In as little as 10 s, for example, peak frequencies shifted by a factor of two. For many cells, these trial-to-trial variations obscured detectable oscillations when all trials were averaged together.
4. The potential role of neuronal maturation in the generation of oscillatory activity was investigated by studying neuronal responses from kittens at 4 wk postnatal. Of the 80 kitten cells studied, 27 exhibited oscillatory discharge. Although oscillations in the kitten visual cortex spanned the same frequency range as that seen in the adult, oscillations in the midfrequency range (36-44 Hz) are more common in the adult cortex.
5. To explore the possibility that oscillations might play a functional role in vision, we investigated the dependence of oscillations on different stimulus parameters. Responses to dichoptically presented drifting gratings showed no relationship between binocular interactions and oscillatory discharge: oscillations were just as likely to occur with non-optimal as with optimal binocular stimuli. For 98 binocular cells which exhibited oscillatory discharge, monocular and binocular responses were compared. 61 of these cells exhibited oscillatory discharge under both monocular and binocular stimulation. The numbers of cells that oscillated solely under binocular stimulation or solely under monocular stimulation were approximately equal (18 and 19, respectively). Among the 61 cells that oscillated with both monocular and binocular stimuli about half showed stronger oscillations for binocular stimulation.
6. To study the role oscillations might play in the integration of cortical responses to stimuli covering multiple receptive fields, we studied the activity of 89 oscillatory cortical cells stimulated by large drifting gratings (coherent stimuli) and flashed bar sequences (incoherent stimuli). The data from these studies suggest that oscillations are not preferentially associated with coherent visual stimuli. 36 of these cells oscillated only with incoherent stimulation, while 22 cells oscillated only with coherent stimulation. For the remaining 31 cells, oscillations typically occurred at higher frequencies for incoherent stimulation than for coherent stimulation.
7. Low contrast drifting gratings were presented to 66 cortical cells to examine whether oscillations are present at very low response levels. Of these cells, 26 displayed some oscillatory activity. For none of these oscillatory cells did the strength of oscillation show any consistent relationship with contrast. When the oscillations observed in low contrast responses are grouped together, we find that the strength of oscillations is inversely related to the mean firing rate.
8. The study of 59 cells in the LGN revealed the presence of oscillatory discharge in 31 cells. Of these cells, 10 exhibited oscillations which were an order of magnitude stronger than those found in the cortex. These strong geniculate oscillations, which are all around 50 Hz in frequency, are clearly independent of visual stimulation: the strength of oscillation was actually stronger when no visual pattern was presented. In contrast to oscillatory cortical cells, these LGN cells exhibited little variation in either oscillation strength or frequency between successive trials.
9. The data are consistent with a model in which oscillations seen in the visual cortex arise from spontaneous oscillations of a sub-population of retinal ganglion cells. Since the physiological characteristics of oscillatory LGN cells are similar to those of some retinal ganglion cells, strong LGN oscillations in the 50 Hz range are likely to originate in the retina. The stimulus independence of both cortical and LGN oscillations is consistent with the proposal that such oscillations reflect spontaneous activity. Simulations based on measured firing properties of retinal ganglion cells and certain assumptions regarding geniculo-cortical connectivity suggest that the observed nature of cortical oscillations in the 50 Hz frequency range can be largely accounted for by the propagation of spontaneous activity through the central visual pathway.
10. The finding of strong and persistent LGN oscillations calls into question the exclusive association of cortical oscillations with intracortical mechanisms. Additionally, the instability and stimulus independence of oscillations support the notion that the oscillatory activity of single cortical neurons does not reflect parameters of patterned visual stimulation and may be an epiphenomenon of no obvious functional significance to the visual system.