Some electrodes in the array yielded weak and inconsistent responses and were excluded from analysis. For Monkey 1, one region of the array did not yield usable signals. The data reported here were obtained from 27 electrodes in Monkey 1 and 62 electrodes in Monkey 2. Each of these electrodes yielded reliable responses and stable estimates of the receptive field centers (SD less than 0.1° across days, mapped by flashing small Gabor stimuli on a rectangular grid that spanned the receptive fields of all the electrodes). Only electrodes with reliable estimates of the receptive field center were used so that we could limit the stimulus period analysis to sites within 0.2 degrees of the stimulus center (see below). This analysis also provided an independent measure of the signal quality recorded from each electrode. In comparison, Nauhaus and colleagues used electrodes for which the time-to-peak versus distance plots were significant (p<0.001), thereby removing all electrodes that failed to provide supporting evidence for traveling waves.
For pre-stimulus STA, any electrode from which at least 25 spikes were obtained was used for analysis. For Monkey 1 this yielded 116 electrodes from 10 recording sessions (23 unique electrodes - many electrodes were recorded on multiple sessions; median spikes per electrode per session: 314, inter-quartile range: 111-778). For Monkey 2 this yielded 655 electrodes from 17 sessions (60 unique electrodes; median spikes: 236, inter-quartile range: 89-596). For stimulus STA, we used electrodes from which at least 25 spikes were obtained, and whose receptive fields were within 0.2 degrees of the center of the stimulus. For Monkeys 1 and 2 this yielded 22 (12 unique; median spikes: 193, inter-quartile range: 75-335) and 46 (31 unique; median spikes: 98, inter-quartile range: 45-185) electrodes. To account for the multiplicity of some electrodes, the STAs from the same electrode were averaged across sessions.
LFP and multiunits were extracted using commercial hardware and software (Blackrock Microsystems). Raw data were filtered between 0.3 Hz (Butterworth filter, 1st order, analog) and 500 Hz (Butterworth, 4th order, digital) to extract the LFP, and digitized at 2 kHz (16 bit resolution). Multiunits were extracted by filtering the raw signal between 250 Hz (Butterworth, 4th order, digital) and 7500 Hz (Butterworth, 3rd order, analog) followed by an amplitude threshold (set at ∼6.25 and ∼4.25 of the signal SD for the two monkeys). To improve the quality of the isolation, the multiunits were further sorted offline (Offline Sorter, Plexon Inc.), although the results were similar when unsorted multiunits were used. LFP was used either raw (no additional filtering) or after filtering between 3-90 or 15-90 Hz (4th order high-pass and 5th order low-pass Butterworth filters; filtered twice in original and time-reversed order to achieve zero phase distortion). LFP from each electrode was independently z-scored. Because the spikes for STA computation were taken at a fixed time relative to stimulus onset, the stimulus locked LFP signal (i.e., the evoked response) was subtracted from each LFP trace before the STA computation.
For pre-stimulus STA, any electrode from which at least 25 spikes were obtained was used for analysis. For Monkey 1 this yielded 116 electrodes from 10 recording sessions (23 unique electrodes - many electrodes were recorded on multiple sessions; median spikes per electrode per session: 314, inter-quartile range: 111-778). For Monkey 2 this yielded 655 electrodes from 17 sessions (60 unique electrodes; median spikes: 236, inter-quartile range: 89-596). For stimulus STA, we used electrodes from which at least 25 spikes were obtained, and whose receptive fields were within 0.2 degrees of the center of the stimulus. For Monkeys 1 and 2 this yielded 22 (12 unique; median spikes: 193, inter-quartile range: 75-335) and 46 (31 unique; median spikes: 98, inter-quartile range: 45-185) electrodes. To account for the multiplicity of some electrodes, the STAs from the same electrode were averaged across sessions.
LFP and multiunits were extracted using commercial hardware and software (Blackrock Microsystems). Raw data were filtered between 0.3 Hz (Butterworth filter, 1st order, analog) and 500 Hz (Butterworth, 4th order, digital) to extract the LFP, and digitized at 2 kHz (16 bit resolution). Multiunits were extracted by filtering the raw signal between 250 Hz (Butterworth, 4th order, digital) and 7500 Hz (Butterworth, 3rd order, analog) followed by an amplitude threshold (set at ∼6.25 and ∼4.25 of the signal SD for the two monkeys). To improve the quality of the isolation, the multiunits were further sorted offline (Offline Sorter, Plexon Inc.), although the results were similar when unsorted multiunits were used. LFP was used either raw (no additional filtering) or after filtering between 3-90 or 15-90 Hz (4th order high-pass and 5th order low-pass Butterworth filters; filtered twice in original and time-reversed order to achieve zero phase distortion). LFP from each electrode was independently z-scored. Because the spikes for STA computation were taken at a fixed time relative to stimulus onset, the stimulus locked LFP signal (i.e., the evoked response) was subtracted from each LFP trace before the STA computation.
