Different origins of gamma rhythm and high-gamma activity in macaque visual cortex.
Ray S, Maunsell JH. · 2011
Read Different origins of gamma rhythm and high-gamma activity in macaque visual cortex. by Ray S, Maunsell JH. (2011) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/europmc-pmc3075230.txt.
Introduction
Ray and Maunsell settled in 2011 a dispute about fast brain rhythms in monkey visual cortex. The dispute was whether the fastest band is a true rhythm or just the blur of neurons firing. Brain waves above the classic gamma range, called high gamma, appear in many recordings and are used as a signal of local activity. The team varied stimulus size, which pushes firing rates and the classic gamma rhythm in opposite directions.
High gamma power tracked the firing of nearby neurons under every manipulation, rising and falling with spike rates rather than with the rhythm. The two signals also varied together trial by trial. The authors conclude high gamma largely reflects local spiking, not a separate communication channel. That is useful news, since the easy to record wave can stand in for hard to record spikes. But studies must not double count the two as separate signals.
Catalog record
| CREATOR | Ray S, Maunsell JH. |
|---|---|
| YEAR | 2011 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 75 KB |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Different Origins of Gamma Rhythm and High-Gamma Activity in Macaque Visual Cortex By Supratim Ray, John H. R. Maunsell, Leslie Ungerleider Published in: PLoS Biology DOI: 10.1371/journal.pbio.1000610 ABSTRACT High-gamma (80–200 Hz) activity can be dissociated from gamma rhythms in the monkey cortex, and appears largely to reflect spiking activity in the vicinity of the electrode. INTRODUCTION Neuronal oscillations exist in the brain over a wide range of frequencies, including the delta (1–3 Hz), theta (4–8 Hz), alpha (9–12 Hz), beta (12–30 Hz), and gamma (30–80 Hz) bands, and are thought to reflect cortical processing [1]. In addition to the modulation of power in some of the aforementioned frequency bands, there is often an increase in power in a broad frequency range above 80 Hz, called the “high-gamma” band (80–200 Hz). This increase in high-gamma power has been most consistently observed in electrocorticogram (ECoG) studies in humans [2]–[5] but is also observed in local field potentials (LFPs; [6]–[8]) and magnetoencephalogram (MEG; [9],[10]). It has been seen in several cortical areas, under diverse stimulus conditions and a range of cognitive states (for a review see [11]). The functional significance of high-gamma activity and its relationship to gamma rhythms that are typically observed at lower frequencies (30–80 Hz) remain unclear. One suggested role of the gamma band is to provide communication channels between cortical areas [12],[13]. Within this framework, there could be multiple frequency bands for communication [14], so the high-gamma band could serve as a distinct channel [15], possibly nested within a low frequency rhythm [3],[5],[16]. On the other hand, several studies have shown that spiking activity is coupled to power in the high-gamma range [6]–[8],[17]–[19]. Because under many conditions the gamma power and firing rates are correlated (for example, during attentional modulation), it is difficult to distinguish between the two possibilities described above. We addressed this issue by studying the LFP power spectrum in V1 of monkeys while manipulating the stimulus size, because increasing the size decreases the firing rate but increases the strength of the gamma rhythm (i.e., the two are anti-correlated; [20]), permitting a dissociation. Using a signal processing technique called Matching Pursuit (MP) that imposes minimal a priori assumptions on LFP decomposition and can simultaneously resolve both transient and oscillatory components in the LFP [8],[18], we studied the relationship between spiking activity and high-gamma power under conditions when the gamma rhythm was absent (no stimulus), weak (small stimulus size), or strong (large size). We found that high-gamma activity was strongly correlated with the multiunit spiking, under different manipulations of stimulus size and temporal frequency. Our results show that multiunit activity can be reliably estimated from the high-gamma power. Further, while investigating the role of high-gamma band in communication or coding, it is important to account for population spiking activity because it may also modulate power in the high-gamma band. RESULTS Recordings were made from an array of 96 electrodes (Blackrock Systems) that was chronically implanted in the right hemisphere of V1 in two monkeys. The receptive fields were in the lower left visual quadrant at an eccentricity of 3–5°. The monkeys performed an orientation change detection task (Figure S1A), where they attended to a Gabor stimulus outside the receptive field while a series of gratings of six different sizes and orientations were presented inside the receptive field of one of the recording sites (new location for each session) for 400 ms with an interstimulus interval of 600 ms (see Materials and Methods for further details). Analysis was restricted to sites whose receptive field centers were within 0.2° from the stimulus center and for which the firing rate was at least 1 spike/s for each of the six sizes. This yielded 15 and 104 sites from Monkeys 1 and 2. Unless stated otherwise, the results shown below were obtained after pooling the data across orientations to increase the statistical power, although similar results were obtained when the analysis was performed only on the preferred orientation. Four analyses were performed. First, we studied the correlation between firing rates and LFP power (as a function of frequency) while varying the stimulus size. For this analysis, firing rates and LFP power were averaged over time (between 200 and 400 ms after stimulus onset) as well as over stimulus repetitions. Second, we computed the correlation between two time-series: firing rates and the average LFP power in different frequency bands, both computed in 2 ms bins and averaged over stimulus repetitions. Third, we computed the trial-by-trial co-variability in firing rates and LFP power in different frequency bands under identical stimulus conditions. Finally, we performed a spike-triggered analysis in two dimensions (time and frequency) to estimate the temporal and spectral components in the LFP that were locked to spikes. Spectral analyses were performed using the MP algorithm (see Materials and Methods for details) and were compared with the more traditional multitaper method [21],[22] in the Supporting Information section. Correlation between Firing Rates and LFP Power as a Function of Frequency Figure 1A shows the average multiunit firing rate of a typical recording site from Monkey 1 when gratings of six different radii (values shown in the inset of Figure 1C) were presented between 0 and 400 ms. The inset shows the firing rate between 200 and 400 ms (thick horizontal black line on the time axis), as a function of stimulus size. As expected, increasing the stimulus size increased the strength of the inhibitory surround, which decreased the firing rate. Figure 1B shows the change in LFP power relative to a baseline period (defined as 0 to 300 ms before stimulus onset) for three different sizes (radii of 0.3°, 1.14°, and 2.4°, shown in Figure S1B). These time-frequency energy difference spectra showed a large broadband increase in power in the first 100 ms after stimulus onset, coinciding with the transient increase in firing rate (Figure 1A). The gamma rhythm, represented by a horizontal band at ∼50 Hz in the time-frequency spectrum, appeared after the initial transient and continued until the stimulus was turned off at 400 ms. Consistent with the results shown in [20], gamma rhythm amplitude increased with increasing stimulus size. We also observed an increase in power over a broad frequency range above the gamma range (>60 Hz). However, power in this band showed the opposite trend—it decreased with increasing stimulus size, similar to the decrease observed in the firing rates. Figure 1C shows the energy between 200 and 400 ms (indicated by thick black lines on the time axes of Figure 1B), as a function of frequency, for the six stimulus sizes (colored traces) as well as the pre-stimulus baseline (black trace). While the power in the gamma range (40–60 Hz, peak at ∼50 Hz) increased with size, beyond the gamma range there was a clear decrease in power with increasing size. Figure 1D–F and G–I show the population average of LFP recordings from all the sites in Monkeys 1 and 2 (15 and 104, respectively). The firing rates were normalized by dividing by the maximum firing rate for each site before averaging (Figure 1D and G). The time-frequency power difference spectra (Figure 1E, H) and the power versus frequency spectra (Figure 1F, I) were averaged across sites on a log scale