Gamma Oscillatory Responses To Global Gratings In Awake Rhesus Monkey V1

Background: Primary Visual Cortex (V1), the first area in the visual cortical processing hierarchy, integrates the input information from LGN and retina and project it to the higher cortical areas. V1 neurons, characterized by their significant orientation selectivity and sharper widths of tuning curves, encodes some basic features effectively.Gamma oscillation observed in this study includes two types: the neuronal gamma oscillation and the Local Field Potential (LFP) gamma band. The neuronal oscillation refers to the progress during which the instant neuronal firing rate oscillates periodically ,i.e. the spike distribution is a longitudinal wave along the time axis. Particularly, neuronal oscillation in the 30-90Hz band is called neuronal gamma oscillation. While the LFP gamma band is the periodic synchronous fluctuation of membrane potential of big mass of neurons (~106-107/mm3) in the 30-90 Hz band. Dipicting the dendritic activity as a sign of inputs into the local area, LFP contributes to spike times in various manners. The gamma oscillation provides temporal reference for higher functions, such as perceptual binding of sensory features into feature, the storage and recall of memory.Methods: The experiment was performed in an awake rhesus monkey. 3 electrode arrays were implanted into V1 cortex of an adult trained rhesus monkey. 8 experimental conditions, i.e 7 orientation angles of the global gratings(-90°, -60°, -30°, 0°, 30°, 60°, 90°) and blank trial, were employed. Both multiunit activity (MUA) and local field potential (LFP) were recorded simultaneously and the monkey's saccade was monitored.Results: (1) Totally 97 units were well isolated through fine spike sorting, among which 73 neurons were significantly orientation-selective. 7 of the 73 neurons depicted strong gamma (~60-70Hz) oscillation in the autocorrelogram only to the global gratings . The power spectrum of the autocorrelogram indicated that the power of this particular neuronal gamma oscillation was orientation- tuned in the same way as their own mean firing rate, that was, the neuronal gamma oscillation tended to be stronger to its perfect orientation angle and nearby angles to which the neuron also fired frequently, and disappeared to the null orientation and in the blank trials. The peak frequencies in the power spectrum of all the gamma-neurons were also tuned and all were tuned in the same way, which was independent of their own orientation selectivity : fastest(~70Hz) in the±90°, slowest(~60Hz) in the 0°.(2) Power spectrum density in 0-100Hz frequency band were estimated and the blank trial was also subtracted as baseline to obtain the change of powerΔP in the frequency domain. The orientation stimulus induced significant power change in both gamma(~60-70Hz), beta(12-28Hz) , theta(5-10Hz) band. The peak frequency in gamma band of LFP from all the 96 electrodes depicted the same tuning pattern: fastest (~70Hz) at the±90°and slowest (~60Hz) at the 0°. This tuning pattern, consistent with the tuning feature of peak frequency of neuronal gamma oscillation but independent of the local neuronal orientation selectivity, formed a V-shaped tuning curve. TheΔP of beta and theta bands were also changed along with orientation but their peak frequencies weren't tuned. Moreover, no neuron was found oscillating at beta or theta bands.(3) For all the electrodes where neurons were detected, the spike-LFP coherences between the spike trains and the LFP in the identical electrode in the 0-100Hz frequency band were estimated to analyze the relationship between neuronal oscillation and LFP. For the electrodes where no gamma neurons were detected, the coherence remained 0.425±0.025 in the whole 0-100Hz band in each experimental condition. For the electrodes where gamma neurons were detected, however, the coherence increased up to 20%-50% (0.5-0.6) strictly at the gamma (60-70Hz) band but remained 0.425±0.025 in other bands. Both the peak value of spike-LFP coherence and the peak frequency were orientation selective. For the gamma neurons , there were 2 consistencies: the individual orientation selectivity of mean firing rates, power of gamma neuronal oscillations and peak value of spike-LFP coherence in gamma band; the global orientation selectivity of the peak frequencies of the LFP power , neuronal gamma oscillations and spike-LFP coherence in the gamma(60– 70 Hz) band. The three peak frequencies were almost identical.Conclusion: The results indicate that the V1 cortical oscillatory responses represents both individual and global features of the neuronal ensemble that encodes the global grating stimulus. (1) The power of neuronal gamma oscillation, besides the mean firing rate, also represents the neuronal individual orientation selectivity, this is the individual feature of the neuronal ensemble; (2) The peak frequencies of neuronal gamma oscillation an LFPs represents the way the neuronal ensemble"understands"the global grating stimulus and it can recognize a pair of symmetry orientation angles. (3) There is a complex interaction between the neuronal gamma oscillation and LFP gamma band. This systematic oscillation must be a representation of the network behavior in the job of encoding the global grating stimulus. It may be suggested that how powerfully a neuron oscillates depends on its individual orientation preference, while how fast it oscillates depends on the ensemble response to the global gratings.