In the last two decades, the direct measure of neural activity from intracranial electrodes implanted in neurosurgical patients for clinical purpose has been increasingly popular among neuroscientists to investigate the neural basis of sensori-motor and cognitive functions. The richness and complexity of the human intracranial recordings approach come partly from the multiplicity of the recorded neural signals. At a macroscopic level of organization, there are two prominent neurophysiological signals. On the one hand, event-related potentials (ERPs; often called local field potentials in iEEG) which are time-locked and largely phase-locked to an event (e.g. a sensory stimulus in the visual modality) and predominant in the lower range of the frequency spectrum (<30 Hz, here referred to as ‘low frequency’ activity, LF). On the other hand, broadband activity which is largely non phase-locked relative to events and typically observed and quantified over a higher frequency range of the spectrum (>30 Hz also known as ‘gamma range’; here ‘high-frequency’ activity, HF). [Authors’ note: LF and HF signals differ both in the frequency range at which they are prevalent, and in terms of their phase-locking relative to the event onset. Event-related responses are (mostly) low-frequency responses while broadband signals are high-frequency mostly non phase-locked responses, distinct from narrow-band gamma oscillations (Ray and Maunsell, 2011; Hermes et al., 2015). For convenience, in keeping with the frequency range in which they are typically observed, we refer to the event-related/evoked response as ‘LF’ and the broadband response as ‘HF’. Note however that there are cases where phase-locking and frequency range may be dissociated, for instance when using high frequency periodic stimulation (e.g. 15 Hz). In this case, parts of the evoked phase-locked response may fall within the typical range of the high-frequency response, such as the 3rd (45 Hz) or 5th (75 Hz) harmonic of the response e.g. Winawer et al., 2013] . A major challenge for the human intracranial approach is to determine which (characteristics) of these signals are most meaningful to understand the neural basis of sensory, cognitive, or motor events. Although early work focused exclusively on characterizing LF activity, the neuroscientific community has now largely shifted its interest to HF signals, which are thought to reflect population-level neuronal firing (Miller et al., 2007; Nir et al., 2007; Ray et al., 2008) and to be more correlated with blood oxygen level dependent (BOLD) activity as recorded with functional magnetic resonance imaging (fMRI) (Mukamel et al., 2005; Hermes et al., 2012; Winawer et al., 2013; Jacques et al., 2016b). Compared to low frequency signals, HF also appears to be more straightforward to characterize the time-course of sensory-motor and cognitive processes (i.e. avoiding the issue of varying polarity and morphology of ERP responses), may be more selective to specific stimuli (Rangarajan et al., 2014), and is typically assumed to reflect more local neural activity (Crone et al., 1998; Miller et al., 2007; Hermes et al., 2012).

However, due to several factors, the degree of validity of these assumptions, and how they can be generalized across complex human brain functions subtended by large-scale neural networks, is largely unknown. First, in a given study, the simultaneous recording of HF and LF often takes place in a single, relatively small and specific region (e.g. the sensorimotor cortex: Crone et al., 1998; Pfurtscheller et al., 2003; Miller et al., 2007; Hermes et al., 2012; the medial occipital cortex: Winawer et al., 2013; or the left posterior superior temporal gyrus: Crone et al., 2001), and often with limited data sets (4–5 individual brains; N=22 in Miller et al., 2007). Second, the two signals are not systematically quantified, with comparisons often limited to determine the number of significant responses for each signal and their overlap (Crone et al., 1998; Crone et al., 2001; Pfurtscheller et al., 2003; Lachaux et al., 2005). Third, studies are often limited in their ability to objectively identify, quantify and compare HF and LF signals with the same analysis parameters (Lachaux et al., 2005; Fisch et al., 2009; Rangarajan et al., 2014; Engell and McCarthy, 2011; Vidal et al., 2010). Altogether, these factors may have led to, or at least substantially enhanced, the typically reported differences between the two types of neurophysiological signals in terms of spatial extent and localization, degree and type of functional selectivity as well as temporal characteristics (Vidal et al., 2010; Winawer et al., 2013; Davidesco et al., 2013; Privman et al., 2011; Nozaradan et al., 2017).

Here, we provide an original contribution to understand the relationship between stimulus-evoked LF and HF neurophysiological activity in the human brain. Two key aspects of our study allow to circumvent the above-mentioned issues. First, we record the neural system for face recognition, a complex and widely distributed function in the human associative cortex (Sergent et al., 1992; Duchaine and Yovel, 2015; Jonas et al., 2016; Grill-Spector et al., 2017), sampled here with a large population (N=121) of implanted individual human brains (>8000 recording sites). Second, we rely on an original frequency-tagging approach (Galloway, 1990; Norcia et al., 2015) providing an objective definition and quantification of simultaneously recorded LF and HF activity in the frequency-domain with a high signal-to-noise ratio and the same analysis parameters (Figure 1).

Figure 1

Download asset Open asset

Overall, we report (1) highly similar spatial patterns of face-selective LF and HF neural activity including relative local spatial extent; (2) an overall lower face-selective HF amplitude in anterior VOTC regions (i.e. the anterior temporal lobe, ATL), which may explain the underevaluation of this region in recent intracranial studies of higher brain function using HF signal only, and (3) strongly shared functional properties (corresponding amplitudes, face-selectivity, onset time) between LF and HF. While these findings point to largely similar properties between the two types of neural signals, LF category-selective activity has a higher signal-to-noise ratio and allows a more extensive exploration of the ATL, implicated in higher cognitive functions, than HF activity. Altogether, these observations clarify the relationship between two prominent neurophysiological functional activities in the human brain, challenging conventional views and the increasing focus on HF activity in human cognitive neuroscience research.

Face-selective activity in the VOTC was identified in the frequency domain following a Fourier transform applied either on the raw SEEG signal, highlighting low frequency (LF) activity as in previous studies with this paradigm (e.g. Jonas et al., 2016; Hagen et al., 2020), or on the time-varying amplitude envelope of the high frequency broadband (HF) signal (Figure 1D). Importantly, responses were quantified at the exact frequency of face stimulation (1.2 Hz) and harmonics (Figure 1E and F). Significant face-selective responses were determined by grouping of the first four harmonics (i.e. summing 1.2, 2.4, 3.6, and 4.8 Hz, Figure 2E) and computing a Z-score transform (z>3.1, p<0.001, Lochy et al., 2018; Jacques et al., 2020). At this statistical threshold, there were 2130 VOTC recording contacts in the gray matter and medial temporal lobe with significant face-selective activity in 118 participants (among 7374 contacts located in gray matter or medial temporal lobe in the VOTC of 121 participants, that is, 28.8% of all recorded contacts).

Figure 2 with 3 supplements see all

Download asset Open asset

Similar macro-scale spatial organization of LF and HF signals in VOTC

To compare the spatial organization of LF and HF signals in VOTC, we examined (1) the spatial distribution and proportions of LF+ and HF+ contacts, as well as (2) the spatial patterns of face-selective amplitude in these two types of contacts across the VOTC.

The spatial distribution of face-selective contacts was determined by examining VOTC maps displaying individual contacts in Talairach space, as well as maps depicting the local proportion of face-selective contacts relative to the number of recorded contacts (Figure 2). These maps indicate that LF+ contacts were widely distributed across the VOTC (Figure 2, Table 1) with a particular focus in a stretch of cortex going from the IOG, through the FG (particularly in its lateral section –latFG- and adjacent OTS) and up to the antFG and surrounding sulci (antOTS and antCoS). In addition, we observed LF+ contacts in the temporal pole (TP) as well as in subcortical structures of the medial temporal lobe (amygdala –AMG- and hippocampus -HIP). HF+ contacts were also distributed across the VOTC with a focus around the same regions as for LF+ contacts (i.e. IOG, FG, antFG, antOTS, antCoS, Figure 2). Overall, the proportions of both LF+ and HF+ contacts showed a gradual reduction from posterior to anterior VOTC (Figure 2C), and were larger in the right compared to the left hemisphere (LF+: RH = 31.7%, 1059/3339 vs. LH = 25.3%, 1021/4035, p<0.0001, two-tailed permutation test; HF+: RH = 9.6%, 322/3339 vs. LH = 8.0%, 322/4035, p<0.01). Notably however, across the VOTC, the proportion of LF+ contacts was significantly higher than HF+ contacts (2080/7374=28.1% for LF+ vs. 644/7374=8.7% for HF+, p<0.0001, two-tailed permutation test, Figure 2B, C).

Computing a map of the relative decrease of proportion of HF+ relative to LF+ contacts (Figure 2B, right) indicates that, while the lower proportion of HF+ contacts was observed throughout the VOTC, it was most evident in the ATL (i.e., anterior to the middle FG). We also found a decrease of proportion of significant contacts in lateral and medial portions of the PTL, apparently producing a more focused distribution of HF+ contacts around the FG region (see below for further analyses of the local spatial extent of both signals).

To further examine the correspondence in the spatial organization of face-selective response across LF and HF signals, we compared the patterns of amplitudes for the two signals across the VOTC. To avoid including noise in the amplitude estimates, LF and HF amplitudes were computed over their respective pools of significant contacts (i.e. LF+ contacts for LF signal and HF+ contacts for HF signal). Amplitude was quantified as the sum of the baseline-subtracted Fourier amplitude over the first 12 harmonics of the face-selective frequency, excluding harmonics of the base stimulation frequency (i.e. 1.2–16.8 Hz excluding 6 and 12 Hz, Figure 1E, F). We first examined the winsorized mean amplitudes (see Materials and methods) across face-selective contacts in each individually-defined anatomical region (Figure 3A). Overall, the patterns of amplitude variations across VOTC regions are largely similar across the two types of signals. For instance, for both signals, the largest amplitudes were recorded in the latFG, followed by the IOG. This similarity in patterns of amplitude is reflected in the strong correlation between the amplitudes computed in each region and hemispheres (Spearman’s Rho = 0.81, 95% C.I.: [0.42–0.99] computed using 17 regions with more than 5 HF+ contacts, Figure 3B, Table 1). In line with this observation, VOTC maps in TAL space (Figure 3C) computed in 12mm x 12mm voxels reveal a striking similarity in the spatial patterns of face-selective response amplitudes for LF and HF signals measured over their respective sets of recording contacts. This similarity manifests in the robust correlation computed between amplitude maps, that is, using mean amplitudes across contacts in each voxel (Pearson’ r on log-transformed data = 0.75, [0.71–0.78], Figure 3D).

Figure 3

Download asset Open asset

Despite the overall similarity in the spatial organization of the amplitude for LF and HF signals, a notable difference was the stronger right hemispheric advantage for LF compared to HF activity in the latFG. Specifically, while face-selective amplitude was consistently larger in the right compared to left hemisphere in the latFG (right hemispheric advantage: 100*(R-L)/R=38.7%, effect-size/cohen’s d=0.58, Figure 3A), this right hemispheric advantage in the latFG was much weaker for HF (right hemispheric advantage = 8.3%; cohen’s d=0.09). Statistical tests of interhemispheric amplitude differences for each signal and anatomical regions were performed using linear mixed model statistics. This revealed that for LF responses in LF+ contacts, face-selective amplitude was significantly larger in the right hemisphere in the latFG (p<0.01, FDR corrected), MTG (p<0.01, FDR corrected) and marginally so in the hippocampus (p=0.064; p=0.013 uncorrected; corrected ps >0.28 for other regions). In contrast, no significant interhemispheric difference was found for the HF signal (all corrected ps >0.22; MTG, antFG, antMTG, AMG, HIP not tested due to insufficient number of HF+ contacts in these regions, see Table 1).

Statistical threshold accounts for the low proportion of HF+ contacts in the ATL

In the previous section, we indicate that the proportion of face-selective HF+ relative to LF+ contacts in the ATL appears disproportionately low compared to the posterior section of VOTC (Figure 2C, D). This observation could result either from a simple effect of statistical threshold linked to a quantitative difference across signal types and VOTC regions, or from a qualitative difference in the relationship between LF and HF signals in the posterior and anterior VOTC.

To address this issue, we first examined the distribution of face-selective Z-scores (i.e. the value used to determine whether a contact shows a significant face-selective response) across VOTC. We computed Z-scores for all recorded VOTC gray matter contacts (i.e. N=7374 contacts) and visualized the mean Z-score along the postero-anterior axis of VOTC (Figure 4A, see Figure 4—figure supplement 1 for full VOTC maps). This revealed two interesting observations: (1) the mean Z-score is overall higher for LF signal (mean across all VOTC contacts = 3.56 +/- 7.31, Figure 4A) compared to HF signal (1.27+/-4.9), and (2) the Z-scores for both LF and HF signals tend to decrease from posterior to anterior VOTC, with an abrupt reduction starting at around TAL coordinate Y = –40 to –30 (Figure 4A). These two observations indicate that a significance threshold of Z>3.1 as used here will lead to a lower proportion of significant contacts specifically for HF signal at Y Talairach coordinates more anterior than –30, roughly corresponding to the posterior border of the ATL. Despite the difference in overall Z-score value across HF and LF signals, the postero-anterior profiles of Z-scores for the two signals were extremely similar (Pearson correlation: r=0.96). In fact, the Pearson correlation computed between LF and HF Z-score profiles was not statistically different from the correlations within signals when using split-halves datasets (1000 random split-halves, two-tailed 95% CI for between = [0.9 0.96]; for within = [0.88 0.98]), suggesting that the correlations between LF and HF profiles were at ceiling. This similarity was best visualized after normalizing the Z-score profiles (i.e. by subtracting the mean and dividing by the standard deviation of each profile, which corresponds to affine transformations of vertical translation and scaling) to minimize distances between profiles (Figure 4B).

Figure 4 with 3 supplements see all

Download asset Open asset

Second, we investigated how changing the statistical threshold affects the spatial distribution of LF+ and HF+ contacts across VOTC by adjusting the Z-score threshold to exactly match the number of significant LF+ (Z>9.48, n=830 contacts) and HF+ (Z>2.32, n=830 contacts) contacts. Under these conditions, the spatial distribution (Figure 4C) and proportion maps (Figure 4D) for LF+ and HF+ contacts were extremely similar, with no meaningful statistical difference in local proportion of contacts across VOTC (two-tailed permutation test, fdr-corrected at alpha = 0.05). This was reflected in the high correlation between the two proportion maps (Spearman Rho = 0.890, 95% C.I. [0.86 0.91]; Figure 4E). The same was true when statistically comparing the postero-anterior proportion profiles for LF+ and HF+ contacts after adjusting the Z-score thresholds (Figure 4F, two-tailed permutation test, fdr-corrected at alpha = 0.05). Pearson correlation confirmed the similarity between the two postero-anterior proportion profiles (r=0.94). This lack of meaningful difference is in stark contrast relative to large differences in the postero-anterior profiles that were computed using the same threshold for LF+ and HF+ contacts (Figure 2C). Similar observations were made when matching the number of LF+ and HF+ contacts using the original Z>3.1 threshold for HF+ contacts and raising the statistical threshold to Z>12.07 for LF+ contacts to equate the number of HF+ and LF+ contacts (Figure 4—figure supplement 2).

To clarify the origin of the SNR difference (computed as the Z-score here) between LF and HF electrophysiological signals, we decomposed the Z-score into its ‘signal’ and ‘noise’ constituents. We addressed (1) the lower Z-score for HF compared to LF and (2) the lower Z-score in the ATL relative to posterior VOTC (Figure 4A). First, to compare signal and noise between LF and HF signals, we collected these responses over all recorded VOTC contacts and computed a ‘signal index’ and a ‘noise index’ (Figure 4—figure supplement 3) that account for the 1/f relationship between EEG amplitude and frequency (Pritchard, 1992; Miller et al., 2009; Podvalny et al., 2015). This revealed that the noise was similar across the two types of signals (0.28+/-0.03 for LF vs. 0.26+/-0.03 for HF, Figure 4—figure supplement 3) but the face-selective signal amplitude was on average almost five times larger for LF compared to HF (mean +/-std: 0.83+/-1.34 for LF vs. 0.15+/-0.45 for HF). This indicates that the lower Z-score for HF compared to LF is driven by a smaller face-selective signal, and not by a higher noise for HF. This is also in line with the observation that LF+ contacts with or without significant HF face-selective response differ mostly in the amplitude of the signal (0.40+/-0.87 for LF+HF vs. 9.32+/-9.13 for LF+ HF+ , Figure 4—figure supplement 3B) rather than the noise (0.77+/-0.28 for LF+ HF vs. 0.77+/-0.22 for LF+ HF+). Second, we investigated the overall lower Z-score in the ATL (excluding MTL) compared to more posterior VOTC regions (Figure 4A). We computed LF and HF signal and noise in three main regions of the VOTC (OCC, PTL, ATL), again using all recorded VOTC contacts (Figure 4—figure supplement 3). This revealed that the mean face-selective signal amplitude within the ATL was 72% (LF) and 92% (HF) smaller than in the PTL. In contrast, the noise in the ATL was only 10% larger (LF) or of equal magnitude (HF) than in the PTL. This indicates that the lower Z-scores in ATL are mostly driven by a smaller face-selective signal amplitude in this region compared to the posterior VOTC.

Altogether, these observations indicate that the apparent disproportionate reduction of HF+ face-selective contacts in the ATL region results from a combination of 3 quantitative factors: (1) the face-selective Z-score for the HF signal is overall lower than for the LF signal; (2) the Z-score for both signals decreases along the posterior-anterior axis of VOTC; (3) face-selective LF+ or HF+ contacts are defined using a common statistical threshold. These observations rule out major qualitative differences in the relationship between signals across the VOTC.

Similar local spatial extent for LF and HF signals

In previous sections, we characterized and compared the spatial distribution of LF and HF signals at a global level across the VOTC. We found a more focused distribution of HF+ contacts around the FG region that could be interpreted as reflecting a narrower spatial distribution of HF responses (Figure 2B). To further explore this issue and characterize the spatial properties of LF and HF signals at a finer scale, we took advantage of the high spatial resolution of SEEG recordings (center-to-center distance between contacts is 3.5 mm, whereas it is commonly 10 mm in ECoG studies). To this end, we measured the variation of signal amplitudes along the length of whole SEEG electrode arrays (i.e. containing 5–15 contacts). To ensure that our analyses included reliable neural responses for both signals in each electrode array, we focused specifically on electrodes arrays containing at least one LF+HF+ contact (N=215 electrode arrays). Electrodes were grouped as a function of the main VOTC region label (i.e. OCC, PTL, ATL) of the recording contact with the maximum amplitude (based on LF signal).

To visualize and compare the spatial extent of LF and HF signals, we investigated the amplitude profiles of the two signals along each electrode array (Figure 5A, Figure 5—figure supplement 1). This revealed that despite the much larger number of significant contacts for LF compared to HF at the threshold of Z>3.1 (Figure 2A, Table 1), a large proportion of electrode arrays exhibited highly similar local amplitude profiles for LF and HF signals (Figure 5A and B; median Pearson correlation coefficient across 215 electrode arrays = 0.69). Yet, there was a wide range of relationships between LF and HF local amplitude profiles (see Figure 5—figure supplements 1–3 for the display of all 215 electrode arrays), with only a small number of profiles showing little similarity (Figure 5A, bottom row, Figure 5B). In addition, as shown on Figure 5C, for 77% of the electrode arrays, the maximum amplitude in both signals where either measured on the very same contact (48% of electrodes) or in the directly adjacent contacts spaced 3.5 mm apart (29% of electrodes). For the remainder of the electrodes, the location of the maximum amplitude was 7 millimeter (11%) or more (12%) apart.

Figure 5 with 5 supplements see all

Download asset Open asset

We quantified the local spatial properties of LF and HF in 2 separate analyses. To ensure that these analyses were performed on the same neural source for LF and HF in each electrode, we only included electrodes in which the recording contacts with the maximal amplitude for LF and HF were located within 3.5 mm of each other (i.e. 77% of the electrodes; similar results were obtained when including 100% of the electrodes).

As a first estimate of the spatial extent, we determined the size of clusters of contiguous significant face-selective contacts separately for each signal and regions. As expected, the size of these clusters was significantly larger for LF than for HF (mean +/-std = 3.83+/-2.66 contacts vs. 2.29+/-1.58 contacts for LF and HF respectively, p<0.0001, two-tailed permutation test; that is, 13.4 mm vs. 8.0 mm for LF and HF), likely owing to the overall higher Z-scores for LF than for HF (Figure 4A).

Second, to quantify the spatial extent of LF and HF signals at a finer scale and, most importantly, independently of the Z-score and statistical threshold, we measured the rate of face-selective amplitude decline as a function of the distance from the maximum amplitude in each electrodes array, which were grouped by main region and hemisphere (Figure 5D). Separately for LF and HF signals, electrodes were first spatially centered with respect to the contact with the largest amplitude. For each signal and electrode, amplitudes at contacts located on both sides of the maximum - and equidistant from it - were then averaged in order to ‘fold’ the electrode around the maximum. The spatial extent for LF and HF signals in each region was estimated by fitting an exponential decay function to the mean amplitude profile (goodness of fit - R² - ranged from 96.9 to 99.7) and locating the distance from the maximum at which the function reached half of its amplitude range. Across regions, the average spatial extent at half-range ranged from 2.3 to 3.0 mm for LF signal and from 1.9 to 3.4 mm for HF signal (Figure 5D). The difference in half-range spatial extent between LF and HF was therefore minimal, with LF being slightly but significantly larger in the left ATL (3.0 vs 2.3 mm for LF and HF respectively, p<0.005, two-tailed permutation test, fdr-corrected). For other regions, differences in spatial extent were not significantly different from each other (−0.5 to 0.5 mm, all ps >0.27). At lower amplitude relative to the peak (i.e. 25% of amplitude range), the spatial extent ranged from 4.8 to 6.0 mm for LF and from 3.8 to 6.8 mm for HF. At this lower relative amplitude, as for the half-range extent, the spatial extent was slightly larger for LF in the left ATL (6.0 vs 4.5 mm for LF and HF, p<0.005, fdr-corrected) but no significant difference was observed for other regions (−0.93 to 1 mm, all ps >0.25). We obtained almost identical observations when restricting HF signal to the high-gamma range, that is 80–160 Hz (Figure 5—figure supplement 4; Figure 5—figure supplement 5).

To sum up, our data highlight two main findings: (1) when using a metric independent of statistical threshold (i.e., rate of amplitude decay), LF and HF have very similar spatial extent, except in the ATL where LF signal is slightly wider; (2) this spatial extent is relatively narrow, with the amplitude being reduced to 50% of the peak amplitude within 2–3.5 mm of the peak and to 25% within 4–7 mm.

Functional and timing correspondence between LF and HF signals

We further explored the functional relationship between LF and HF category-selective signals by characterizing their respective amplitudes, selectivity and timing.

First, we characterized the functional relationship between LF and HF signals by correlating the face-selective amplitudes across signals using single recording contacts as datapoints. To avoid including noise in correlation estimates, we restricted the analyses to the set of LF+HF+ contacts in which a reliable face-selective activity could be measured for both signals. This revealed a strong relationship when considering responses across the VOTC (Pearson’s r on log-transformed amplitudes across 569 contacts: 0.59 [0.53–0.64], p<0.001). Mapping Pearson correlations in Talairach space across the VOTC (Figure 6A) revealed that the highest correlations were found in the posterior VOTC, with local peaks in the right latFG and VMO (r~=0.8–0.9), while they were slightly lower in the anterior VOTC. This was confirmed when computing correlations between LF and HF signals separately for each main VOTC region using individual participants’ anatomical labels (Figure 6B). This revealed similar correlations in the OCC (Pearson’s r on log-transformed amplitudes: 0.63 [0.5–0.73]) and PTL (Pearson’s r: 0.68 [0.61–0.75]) but a slightly lower LF-HF correlation in the ATL (Pearson’s r: 0.45 [0.32–0.58]; OCC vs ATL: p=0.031; PTL vs. ATL: p<0.001, two-tailed percentile bootstrap). Interestingly, within the PTL, the correlation between LF and HF signals was much stronger in the latFG (Pearson’s r: 0.73 [0.63–0.80], log-transformed data) than in the medFG (Pearson’s r: 0.42 [0.25–0.58]).

Figure 6

Download asset Open asset

The interpretation of the magnitude of a correlation depends on the maximum correlation that can be expected as a function of the noise in the data. Thus, to further characterize the strength of the relationship between LF and HF, we computed an estimation of the maximum expected correlation (MEC) in this dataset (Figure 6C), which also relates to test-retest reliability. The noise used to compute the MEC was quantified as the standard deviation in face-selective response amplitude between FPVS sequences. The MEC (averaged across LF and HF) at the level of the whole VOTC was 0.88, 99% CI: [0.85–0.91]. This means that the correlation between LF and HF signals reaches 71%, 99% CI: [67-76]% ( ratio of the actual LF vs. HF correlation to the MEC: 0.59/0.83) of the maximum expected correlation in this dataset. For the main regions of the VOTC, the LF vs. HF correlations reached 74% (OCC: 0.63/0.86; 99% CI: [67-87]%), 78% (PTL: 0.68/0.88; 99% CI: [73-87]%), and 60% (ATL: 0.45/0.76; 99% CI: [54-71]%) of the maximum expected correlation.

Second, we compared the magnitude of face-selectivity across LF and HF signals using a face-selectivity index (FSI) computed by taking the ratio of the face-selective amplitude (i.e. at 1.2 Hz and harmonics) to the sum of the face-selective and general visual (i.e. amplitude at 6, 12, 18 Hz) responses. The FSI varies from 0 (no face-selective response) to 1 (only face-selective response and no general visual response) and allows to quantify the magnitude of the face-selective response relative to the overall visual responsiveness of the cortex around the recording contact. Unlike the face-selective amplitude, which cannot be directly compared across LF and HF, the FSI allows for a direct comparison of selectivity across the two signals. Quantifying the FSI (using LF+HF+ contacts) reveals three important findings. First, the face-selectivity indices were virtually identical between LF and HF signals. Comparing LF to HF FSI for each main VOTC region and hemisphere (Figure 7A) revealed no significant difference between signals (range of FSI difference: –0.023–0.041; all ps >0.6, fdr-corrected). Second, selectivity indices increased from posterior VOTC / OCC (LF = 0.64, 99% confidence interval: [0.60–0.69]; HF = 0.62 [0.56–0.67]) to anterior VOTC / ATL (LF = 0.87 [0.85–0.89] and HF = 0.86 [0.84–0.89]) (Figure 7B). Third, the face-selectivity indices were larger in the right than in the left hemisphere in OCC (averaging LF and HF: left = 0.57 [0.5–0.64] vs. right = 0.71 [0.65–0.76], p<0.005, fdr-corrected) and PTL (left = 0.74 [0.7–0.77] vs. right = 0.80 [0.76–0.83], p<0.005) regions. In the ATL, the FSI was slightly larger in the left hemisphere (left = 0.89 [0.87–0.91] vs. right = 0.85 [0.82–0.88], p<0.01). Computing a VOTC map of the face-selectivity index in the Talairach space corroborated these three observations (Figure 7C). The maps further reveal a low face-selectivity index in medial VOTC for both LF and HF signals, and a high index over more lateral regions, in the strip of cortex from IOG to ATL, where the largest face-selective amplitude is measured. However, unlike face-selective amplitude, the largest face-selectivity index is measured in the ATL regions, owing to the very low response to non-face stimuli (i.e. captured by the 6 Hz general visual response). This means that when taking into account the amplitude of the neural response to various visual categories, face-selectivity is similar across the two types of signals.

Figure 7

Download asset Open asset

Third, we explored the time course of LF and HF signals by focusing on the VOTC region with the largest response for both signals: the right latFG. Face-selective time-domain responses were obtained at each contact by selectively filtering-out the SEEG signal generated by the non-face stimuli presented at 6 Hz and harmonics (Retter and Rossion, 2016; Retter et al., 2020), and segmenting the FPVS time-series around the onset of each face in the FPVS sequences. Figure 8A shows normalized LF and HF time-domain face-selective responses in three individual recording contacts with either a high Z-score (left and middle plots), or a middle/low Z-score (right plot). These plots illustrate variations in voltage polarity for LF (mostly positive in the left plot or negative in the middle plot) and variability in SNR across recording contacts (i.e. noisier on the right).

Figure 8

Download asset Open asset

To investigate the correspondence between LF and HF face-selective responses in the latFG we visualized the time-domain response averaged across LF+ HF+ recording contacts (N=65). As shown in Figure 8B, both signals started to deviate from baseline at around the same latency (LF: 79ms, 95% confidence interval: [58 - 101] ms and HF: 91ms, [82 - 103] ms), although the LF signal significantly rose above baseline slightly later than HF (102ms, [96 - 108] ms vs. 77ms, [75 - 81] ms for LF and HF respectively) due to higher across-trial standard deviation for LF. In addition, the response duration was longer for LF, returning to baseline level at 639ms (95% confidence interval: [637 641]), compared to HF signal which returned to baseline level at 409 ms [402 - 430].

To further characterize the timing relationship across signals, we estimated the onset latency of LF and HF face-selective response in each right latFG recording contact (rejecting 4 contacts in which we could not reliably determine onset latency). This revealed a highly significant correlation between LF and HF onset latencies (Pearson’s r=0.48, p<0.001, 99% CI: [0.23 0.67], Figure 8C). Despite a significant correlation, the magnitude of the correlation suffers from the uncertainty in the estimated precise onset latency in a number of contacts with lower SNR. Hence, the correlation across latencies was higher when including only contacts with higher Z-score. For instance, Pearson’s r was 0.63 when only including the 30 contacts with highest Z-score and peaked at r=0.68 for the 16 contacts with highest Z-scores.

Here, we provide a large-scale comparison between two widely used electrophysiological markers of neural activity to investigate human brain function in Neuroscience research: low-frequency signals, that are both time-locked and phase-locked to the stimulation, and high frequency broadband signals, which are time-locked but largely non phase-locked to the stimulation. To this end, we analyzed an unusually large dataset of 121 participants providing more than 8000 recording sites across the whole VOTC. We measured neural activity selectively triggered by the presentation of human faces, an optimal stimulus category to probe the ventral pathway for visual object recognition (DiCarlo et al., 2012; Grill-Spector et al., 2017; Rossion et al., 2018). Our original approach combines (1) intracerebral recordings, which give access to both sulci and gyri, and (2) frequency-tagging to maximize homogeneity across the two signal analyses pipelines, signal-to-noise ratio, and objectivity of measurement of both signal and noise (i.e., at pre-defined frequency bins). It reveals three main findings overall.

First, we find highly similar spatial patterns of LF and HF face-selective activity across brain regions, in terms of significant contacts overlap, relative amplitudes and local spatial extent. Second, our analyses show that genuine category-selective neural activity in the anterior portion of the VOTC, the VATL, is likely to be missed with HF signals only. Finally, these two face-selective neural signals are highly functionally related as indicated by their strong amplitude correlation, degree of face-selectivity, and concurrent onset timing.

Overall, these findings point to largely similar functional properties between these two major neural signals, which may, at the current state of knowledge, therefore provide essentially the same type of information regarding the neural basis of human (re)cognition. Yet, questioning the current focus on HF signals at the expense of LF signals in intracranial human recording research, our observations highlight significant advantages of LF signals: higher SNR to identify a larger number of significant responses, stronger right hemispheric dominance and more extensive ATL exploration. Altogether, these observations, which are discussed more specifically below, re-establish the high value of LF in mapping neural activity in human associative cortex with iEEG.

All SEEG data generated or analysed during this study have been uploaded to Dyad (https://doi.org/10.5061/dryad.66t1g1k51).

1. 1. Rossion B

   (2022) Dryad Digital Repository

   Data from: Low and high frequency intracranial neural signals match in the human associative cortex.

   https://doi.org/10.5061/dryad.66t1g1k51

1. 1. Adrian ED
   2. Matthews BHC

   (1934) THE berger rhythm: potential changes from the occipital lobes in man

   Brain 57:355–385.

   https://doi.org/10.1093/brain/57.4.355

   • Google Scholar
2. 1. Allison T
   2. Puce A
   3. Spencer DD
   4. McCarthy G
   5. Belger A

   (1999) Electrophysiological studies of human face perception; I: potentials generated in occipitotemporal cortex by face and non-face stimuli

   Cerebral Cortex 9:415–430.

   https://doi.org/10.1093/cercor/9.5.415

   • PubMed
   • Google Scholar
3. 1. Bates D
   2. Mächler M
   3. Bolker B
   4. Walker S

   (2015) Fitting linear mixed-effects models using lme4

   Journal of Statistical Software 67:1–48.

   https://doi.org/10.18637/jss.v067.i01

   • Google Scholar
4. 1. Bédos Ulvin L
   2. Jonas J
   3. Brissart H
   4. Colnat-Coulbois S
   5. Thiriaux A
   6. Vignal J-P
   7. Maillard L

   (2017) Intracerebral stimulation of left and right ventral temporal cortex during object naming

   Brain and Language 175:71–76.

   https://doi.org/10.1016/j.bandl.2017.09.003

   • PubMed
   • Google Scholar
5. 1. Benjamini Y
   2. Hochberg Y

   (1995) Controlling the false discovery rate: A practical and powerful approach to multiple testing

   Journal of the Royal Statistical Society 57:289–300.

   https://doi.org/10.1111/j.2517-6161.1995.tb02031.x

   • Google Scholar
6. Book

   1. Bryant KL
   2. Preuss TM

   (2018) A comparative perspective on the human temporal lobe

   In: Bruner E, Ogihara N, Tanabe HC, editors. Digital Endocasts: From Skulls to Brains. Tokyo: Springer Japan. pp. 239–258.

   https://doi.org/10.1007/978-4-431-56582-6_16

   • PubMed
   • Google Scholar
7. 1. Buzsáki G

   (2004) Large-scale recording of neuronal ensembles

   Nature Neuroscience 7:446–451.

   https://doi.org/10.1038/nn1233

   • PubMed
   • Google Scholar
8. 1. Buzsáki G
   2. Anastassiou CA
   3. Koch C

   (2012) The origin of extracellular fields and currents--EEG, ecog, LFP and spikes

   Nature Reviews. Neuroscience 13:407–420.

   https://doi.org/10.1038/nrn3241

   • PubMed
   • Google Scholar
9. 1. Collins CE
   2. Airey DC
   3. Young NA
   4. Leitch DB
   5. Kaas JH

   (2010) Neuron densities vary across and within cortical areas in primates

   PNAS 107:15927–15932.

   https://doi.org/10.1073/pnas.1010356107

   • PubMed
   • Google Scholar
10. 1. Collins JA
    2. Olson IR

    (2014) Beyond the FFA: the role of the ventral anterior temporal lobes in face processing

    Neuropsychologia 61:65–79.

    https://doi.org/10.1016/j.neuropsychologia.2014.06.005

    • PubMed
    • Google Scholar
11. 1. Cox RW

    (1996) AFNI: software for analysis and visualization of functional magnetic resonance neuroimages

    Computers and Biomedical Research, an International Journal 29:162–173.

    https://doi.org/10.1006/cbmr.1996.0014

    • PubMed
    • Google Scholar
12. 1. Crone NE
    2. Miglioretti DL
    3. Gordon B
    4. Lesser RP

    (1998) Functional mapping of human sensorimotor cortex with electrocorticographic spectral analysis II event-related synchronization in the gamma band

    Brain: A Journal of Neurology 121:2301–2315.

    https://doi.org/10.1093/brain/121.12.2301

    • PubMed
    • Google Scholar
13. 1. Crone NE
    2. Boatman D
    3. Gordon B
    4. Hao L

    (2001) Induced electrocorticographic gamma activity during auditory perception: brazier award-winning article, 2001

    Clinical Neurophysiology 112:565–582.

    https://doi.org/10.1016/s1388-2457(00)00545-9

    • PubMed
    • Google Scholar
14. 1. Crone NE
    2. Sinai A
    3. Korzeniewska A

    (2006) High-frequency gamma oscillations and human brain mapping with electrocorticography

    Progress in Brain Research 159:275–295.

    https://doi.org/10.1016/S0079-6123(06)59019-3

    • PubMed
    • Google Scholar
15. 1. Davidesco I
    2. Harel M
    3. Ramot M
    4. Kramer U
    5. Kipervasser S
    6. Andelman F
    7. Neufeld MY
    8. Goelman G
    9. Fried I
    10. Malach R

    (2013) Spatial and object-based attention modulates broadband high-frequency responses across the human visual cortical hierarchy

    The Journal of Neuroscience 33:1228–1240.

    https://doi.org/10.1523/JNEUROSCI.3181-12.2013

    • PubMed
    • Google Scholar
16. 1. DiCarlo JJ
    2. Zoccolan D
    3. Rust NC

    (2012) How does the brain solve visual object recognition?

    Neuron 73:415–434.

    https://doi.org/10.1016/j.neuron.2012.01.010

    • PubMed
    • Google Scholar
17. 1. Dixon WJ

    (1960) Simplified estimation from censored normal samples

    The Annals of Mathematical Statistics 31:385–391.

    https://doi.org/10.1214/aoms/1177705900

    • Google Scholar
18. 1. Dubey A
    2. Ray S

    (2019) Cortical electrocorticogram (ecog) is a local signal

    The Journal of Neuroscience 39:4299–4311.

    https://doi.org/10.1523/JNEUROSCI.2917-18.2019

    • PubMed
    • Google Scholar
19. 1. Duchaine B
    2. Yovel G

    (2015) A revised neural framework for face processing

    Annual Review of Vision Science 1:393–416.

    https://doi.org/10.1146/annurev-vision-082114-035518

    • PubMed
    • Google Scholar
20. 1. Engell AD
    2. McCarthy G

    (2011) The relationship of γ oscillations and face-specific erps recorded subdurally from occipitotemporal cortex

    Cerebral Cortex 21:1213–1221.

    https://doi.org/10.1093/cercor/bhq206

    • PubMed
    • Google Scholar
21. 1. Fisch L
    2. Privman E
    3. Ramot M
    4. Harel M
    5. Nir Y
    6. Kipervasser S
    7. Andelman F
    8. Neufeld MY
    9. Kramer U
    10. Fried I
    11. Malach R

    (2009) Neural “ignition”: enhanced activation linked to perceptual awareness in human ventral stream visual cortex

    Neuron 64:562–574.

    https://doi.org/10.1016/j.neuron.2009.11.001

    • PubMed
    • Google Scholar
22. 1. Galloway NR

    (1990) Human brain electrophysiology: evoked potentials and evoked magnetic fields in science and medicine

    British Journal of Ophthalmology 74:255.

    https://doi.org/10.1136/bjo.74.4.255-a

    • Google Scholar
23. 1. Gao X
    2. Gentile F
    3. Rossion B

    (2018) Fast periodic stimulation (FPS): a highly effective approach in fmri brain mapping

    Brain Structure & Function 223:2433–2454.

    https://doi.org/10.1007/s00429-018-1630-4

    • PubMed
    • Google Scholar
24. 1. Gray CM
    2. Maldonado PE
    3. Wilson M
    4. McNaughton B

    (1995) Tetrodes markedly improve the reliability and yield of multiple single-unit isolation from multi-unit recordings in cat striate cortex

    Journal of Neuroscience Methods 63:43–54.

    https://doi.org/10.1016/0165-0270(95)00085-2

    • PubMed
    • Google Scholar
25. 1. Grill-Spector K
    2. Weiner KS
    3. Kay KN
    4. Gomez J

    (2017) The functional neuroanatomy of human face perception

    Annual Review of Vision Science 3:167–196.

    https://doi.org/10.1146/annurev-vision-102016-061214

    • PubMed
    • Google Scholar
26. 1. Hagen S
    2. Jacques C
    3. Maillard L
    4. Colnat-Coulbois S
    5. Rossion B
    6. Jonas J

    (2020) Spatially dissociated intracerebral maps for face- and house-selective activity in the human ventral occipito-temporal cortex

    Cerebral Cortex 30:4026–4043.

    https://doi.org/10.1093/cercor/bhaa022

    • PubMed
    • Google Scholar
27. 1. Haxby JV
    2. Hoffman EA
    3. Gobbini MI

    (2000) The distributed human neural system for face perception

    Trends in Cognitive Sciences 4:223–233.

    https://doi.org/10.1016/s1364-6613(00)01482-0

    • PubMed
    • Google Scholar
28. 1. Hermes D
    2. Miller KJ
    3. Vansteensel MJ
    4. Aarnoutse EJ
    5. Leijten FSS
    6. Ramsey NF

    (2012) Neurophysiologic correlates of fmri in human motor cortex

    Human Brain Mapping 33:1689–1699.

    https://doi.org/10.1002/hbm.21314

    • PubMed
    • Google Scholar
29. 1. Hermes D
    2. Miller KJ
    3. Wandell BA
    4. Winawer J

    (2015) Stimulus dependence of gamma oscillations in human visual cortex

    Cerebral Cortex 25:2951–2959.

    https://doi.org/10.1093/cercor/bhu091

    • PubMed
    • Google Scholar
30. 1. Herreras O

    (2016) Local field potentials: myths and misunderstandings

    Frontiers in Neural Circuits 10:101.

    https://doi.org/10.3389/fncir.2016.00101

    • PubMed
    • Google Scholar
31. 1. Jacques C
    2. Retter TL
    3. Rossion B

    (2016a) A single glance at natural face images generate larger and qualitatively different category-selective spatio-temporal signatures than other ecologically-relevant categories in the human brain

    NeuroImage 137:21–33.

    https://doi.org/10.1016/j.neuroimage.2016.04.045

    • PubMed
    • Google Scholar
32. 1. Jacques C
    2. Witthoft N
    3. Weiner KS
    4. Foster BL
    5. Rangarajan V
    6. Hermes D
    7. Miller KJ
    8. Parvizi J
    9. Grill-Spector K

    (2016b) Corresponding ecog and fmri category-selective signals in human ventral temporal cortex

    Neuropsychologia 83:14–28.

    https://doi.org/10.1016/j.neuropsychologia.2015.07.024

    • PubMed
    • Google Scholar
33. 1. Jacques C
    2. Rossion B
    3. Volfart A
    4. Brissart H
    5. Colnat-Coulbois S
    6. Maillard L
    7. Jonas J

    (2020) The neural basis of rapid unfamiliar face individuation with human intracerebral recordings

    NeuroImage 221:117174.

    https://doi.org/10.1016/j.neuroimage.2020.117174

    • PubMed
    • Google Scholar
34. 1. Jonas J
    2. Descoins M
    3. Koessler L
    4. Colnat-Coulbois S
    5. Sauvée M
    6. Guye M
    7. Vignal JP
    8. Vespignani H
    9. Rossion B
    10. Maillard L

    (2012) Focal electrical intracerebral stimulation of a face-sensitive area causes transient prosopagnosia

    Neuroscience 222:281–288.

    https://doi.org/10.1016/j.neuroscience.2012.07.021

    • PubMed
    • Google Scholar
35. 1. Jonas J.
    2. Rossion B
    3. Brissart H
    4. Frismand S
    5. Jacques C
    6. Hossu G
    7. Colnat-Coulbois S
    8. Vespignani H
    9. Vignal JP
    10. Maillard L

    (2015) Beyond the core face-processing network: intracerebral stimulation of a face-selective area in the right anterior fusiform gyrus elicits transient prosopagnosia

    Cortex; a Journal Devoted to the Study of the Nervous System and Behavior 72:140–155.

    https://doi.org/10.1016/j.cortex.2015.05.026

    • PubMed
    • Google Scholar
36. 1. Jonas J
    2. Jacques C
    3. Liu-Shuang J
    4. Brissart H
    5. Colnat-Coulbois S
    6. Maillard L
    7. Rossion B

    (2016) A face-selective ventral occipito-temporal map of the human brain with intracerebral potentials

    PNAS 113:E4088–E4097.

    https://doi.org/10.1073/pnas.1522033113

    • PubMed
    • Google Scholar
37. 1. Jonas J
    2. Brissart H
    3. Hossu G
    4. Colnat-Coulbois S
    5. Vignal JP
    6. Rossion B
    7. Maillard L

    (2018) A face identity hallucination (palinopsia) generated by intracerebral stimulation of the face-selective right lateral fusiform cortex

    Cortex; a Journal Devoted to the Study of the Nervous System and Behavior 99:296–310.

    https://doi.org/10.1016/j.cortex.2017.11.022

    • PubMed
    • Google Scholar
38. 1. Kadipasaoglu CM
    2. Conner CR
    3. Whaley ML
    4. Baboyan VG
    5. Tandon N

    (2016) Category-selectivity in human visual cortex follows cortical topology: A grouped iceeg study

    PLOS ONE 11:e0157109.

    https://doi.org/10.1371/journal.pone.0157109

    • PubMed
    • Google Scholar
39. 1. Kay KN
    2. Winawer J
    3. Mezer A
    4. Wandell BA

    (2013) Compressive spatial summation in human visual cortex

    Journal of Neurophysiology 110:481–494.

    https://doi.org/10.1152/jn.00105.2013

    • PubMed
    • Google Scholar
40. 1. Kim JJ
    2. Crespo-Facorro B
    3. Andreasen NC
    4. O’Leary DS
    5. Zhang B
    6. Harris G
    7. Magnotta VA

    (2000) An MRI-based parcellation method for the temporal lobe

    NeuroImage 11:271–288.

    https://doi.org/10.1006/nimg.2000.0543

    • PubMed
    • Google Scholar
41. 1. Kovács G

    (2020) Getting to know someone: familiarity, person recognition, and identification in the human brain

    Journal of Cognitive Neuroscience 32:2205–2225.

    https://doi.org/10.1162/jocn_a_01627

    • PubMed
    • Google Scholar
42. 1. Lachaux JP
    2. George N
    3. Tallon-Baudry C
    4. Martinerie J
    5. Hugueville L
    6. Minotti L
    7. Kahane P
    8. Renault B

    (2005) The many faces of the gamma band response to complex visual stimuli

    NeuroImage 25:491–501.

    https://doi.org/10.1016/j.neuroimage.2004.11.052

    • PubMed
    • Google Scholar
43. 1. Lachaux JP
    2. Axmacher N
    3. Mormann F
    4. Halgren E
    5. Crone NE

    (2012) High-frequency neural activity and human cognition: past, present and possible future of intracranial EEG research

    Progress in Neurobiology 98:279–301.

    https://doi.org/10.1016/j.pneurobio.2012.06.008

    • PubMed
    • Google Scholar
44. 1. Lambon Ralph MA
    2. Sage K
    3. Jones RW
    4. Mayberry EJ

    (2010) Coherent concepts are computed in the anterior temporal lobes

    PNAS 107:2717–2722.

    https://doi.org/10.1073/pnas.0907307107

    • PubMed
    • Google Scholar
45. 1. Leszczyński M
    2. Barczak A
    3. Kajikawa Y
    4. Ulbert I
    5. Falchier AY
    6. Tal I
    7. Haegens S
    8. Melloni L
    9. Knight RT
    10. Schroeder CE

    (2020) Dissociation of broadband high-frequency activity and neuronal firing in the neocortex

    Science Advances 6:eabb0977.

    https://doi.org/10.1126/sciadv.abb0977

    • PubMed
    • Google Scholar
46. 1. Li G
    2. Jiang S
    3. Paraskevopoulou SE
    4. Wang M
    5. Xu Y
    6. Wu Z
    7. Chen L
    8. Zhang D
    9. Schalk G

    (2018) Optimal referencing for stereo-electroencephalographic (SEEG) recordings

    NeuroImage 183:327–335.

    https://doi.org/10.1016/j.neuroimage.2018.08.020

    • PubMed
    • Google Scholar
47. 1. Lindén H
    2. Tetzlaff T
    3. Potjans TC
    4. Pettersen KH
    5. Grün S
    6. Diesmann M
    7. Einevoll GT

    (2011) Modeling the spatial reach of the LFP

    Neuron 72:859–872.

    https://doi.org/10.1016/j.neuron.2011.11.006

    • PubMed
    • Google Scholar
48. 1. Lochy A
    2. Jacques C
    3. Maillard L
    4. Colnat-Coulbois S
    5. Rossion B
    6. Jonas J

    (2018) Selective visual representation of letters and words in the left ventral occipito-temporal cortex with intracerebral recordings

    PNAS 115:E7595–E7604.

    https://doi.org/10.1073/pnas.1718987115

    • PubMed
    • Google Scholar
49. 1. Manning JR
    2. Jacobs J
    3. Fried I
    4. Kahana MJ

    (2009) Broadband shifts in local field potential power spectra are correlated with single-neuron spiking in humans

    The Journal of Neuroscience 29:13613–13620.

    https://doi.org/10.1523/JNEUROSCI.2041-09.2009

    • PubMed
    • Google Scholar
50. 1. Miller KJ
    2. Leuthardt EC
    3. Schalk G
    4. Rao RPN
    5. Anderson NR
    6. Moran DW
    7. Miller JW
    8. Ojemann JG

    (2007) Spectral changes in cortical surface potentials during motor movement

    The Journal of Neuroscience 27:2424–2432.

    https://doi.org/10.1523/JNEUROSCI.3886-06.2007

    • PubMed
    • Google Scholar
51. 1. Miller KJ
    2. Sorensen LB
    3. Ojemann JG
    4. den Nijs M

    (2009) Power-law scaling in the brain surface electric potential

    PLOS Computational Biology 5:e1000609.

    https://doi.org/10.1371/journal.pcbi.1000609

    • PubMed
    • Google Scholar
52. 1. Miller KJ
    2. Schalk G
    3. Hermes D
    4. Ojemann JG
    5. Rao RPN

    (2016) Spontaneous decoding of the timing and content of human object perception from cortical surface recordings reveals complementary information in the event-related potential and broadband spectral change

    PLOS Computational Biology 12:e1004660.

    https://doi.org/10.1371/journal.pcbi.1004660

    • PubMed
    • Google Scholar
53. 1. Miller KJ
    2. Hermes D
    3. Pestilli F
    4. Wig GS
    5. Ojemann JG

    (2017) Face percept formation in human ventral temporal cortex

    Journal of Neurophysiology 118:2614–2627.

    https://doi.org/10.1152/jn.00113.2017

    • PubMed
    • Google Scholar
54. 1. Mukamel R
    2. Gelbard H
    3. Arieli A
    4. Hasson U
    5. Fried I
    6. Malach R

    (2005) Coupling between neuronal firing, field potentials, and FMRI in human auditory cortex

    Science 309:951–954.

    https://doi.org/10.1126/science.1110913

    • PubMed
    • Google Scholar
55. 1. Nir Y
    2. Fisch L
    3. Mukamel R
    4. Gelbard-Sagiv H
    5. Arieli A
    6. Fried I
    7. Malach R

    (2007) Coupling between neuronal firing rate, gamma LFP, and BOLD fmri is related to interneuronal correlations

    Current Biology 17:1275–1285.

    https://doi.org/10.1016/j.cub.2007.06.066

    • PubMed
    • Google Scholar
56. 1. Norcia AM
    2. Appelbaum LG
    3. Ales JM
    4. Cottereau BR
    5. Rossion B

    (2015) The steady-state visual evoked potential in vision research: A review

    Journal of Vision 15:4.

    https://doi.org/10.1167/15.6.4

    • PubMed
    • Google Scholar
57. 1. Norman Y
    2. Yeagle EM
    3. Khuvis S
    4. Harel M
    5. Mehta AD
    6. Malach R

    (2019) Hippocampal sharp-wave ripples linked to visual episodic recollection in humans

    Science 365:eaax1030.

    https://doi.org/10.1126/science.aax1030

    • PubMed
    • Google Scholar
58. 1. Nozaradan S
    2. Mouraux A
    3. Jonas J
    4. Colnat-Coulbois S
    5. Rossion B
    6. Maillard L

    (2017) Intracerebral evidence of rhythm transform in the human auditory cortex

    Brain Structure & Function 222:2389–2404.

    https://doi.org/10.1007/s00429-016-1348-0

    • PubMed
    • Google Scholar
59. 1. Ojemann JG
    2. Akbudak E
    3. Snyder AZ
    4. McKinstry RC
    5. Raichle ME
    6. Conturo TE

    (1997) Anatomic localization and quantitative analysis of gradient refocused echo-planar fmri susceptibility artifacts

    NeuroImage 6:156–167.

    https://doi.org/10.1006/nimg.1997.0289

    • PubMed
    • Google Scholar
60. 1. Persichetti AS
    2. Denning JM
    3. Gotts SJ
    4. Martin A

    (2021) A data-driven functional mapping of the anterior temporal lobes

    The Journal of Neuroscience 41:6038–6049.

    https://doi.org/10.1523/JNEUROSCI.0456-21.2021

    • PubMed
    • Google Scholar
61. 1. Pfurtscheller G
    2. Lopes da Silva FH

    (1999) Event-related EEG/MEG synchronization and desynchronization: basic principles

    Clinical Neurophysiology 110:1842–1857.

    https://doi.org/10.1016/s1388-2457(99)00141-8

    • PubMed
    • Google Scholar
62. 1. Pfurtscheller G
    2. Graimann B
    3. Huggins JE
    4. Levine SP
    5. Schuh LA

    (2003) Spatiotemporal patterns of beta desynchronization and gamma synchronization in corticographic data during self-paced movement

    Clinical Neurophysiology 114:1226–1236.

    https://doi.org/10.1016/s1388-2457(03)00067-1

    • PubMed
    • Google Scholar
63. 1. Podvalny E
    2. Noy N
    3. Harel M
    4. Bickel S
    5. Chechik G
    6. Schroeder CE
    7. Mehta AD
    8. Tsodyks M
    9. Malach R

    (2015) A unifying principle underlying the extracellular field potential spectral responses in the human cortex

    Journal of Neurophysiology 114:505–519.

    https://doi.org/10.1152/jn.00943.2014

    • PubMed
    • Google Scholar
64. 1. Potter MC
    2. Levy EI

    (1969) Recognition memory for a rapid sequence of pictures

    Journal of Experimental Psychology 81:10–15.

    https://doi.org/10.1037/h0027470

    • PubMed
    • Google Scholar
65. 1. Potter MC
    2. Wyble B
    3. Hagmann CE
    4. McCourt ES

    (2014) Detecting meaning in RSVP at 13 ms per picture

    Attention, Perception & Psychophysics 76:270–279.

    https://doi.org/10.3758/s13414-013-0605-z

    • PubMed
    • Google Scholar
66. 1. Pritchard WS

    (1992) The brain in fractal time: 1/f-like power spectrum scaling of the human electroencephalogram

    The International Journal of Neuroscience 66:119–129.

    https://doi.org/10.3109/00207459208999796

    • PubMed
    • Google Scholar
67. 1. Privman E
    2. Fisch L
    3. Neufeld MY
    4. Kramer U
    5. Kipervasser S
    6. Andelman F
    7. Yeshurun Y
    8. Fried I
    9. Malach R

    (2011) Antagonistic relationship between gamma power and visual evoked potentials revealed in human visual cortex

    Cerebral Cortex 21:616–624.

    https://doi.org/10.1093/cercor/bhq128

    • PubMed
    • Google Scholar
68. 1. Quek GL
    2. Rossion B

    (2017) Category-selective human brain processes elicited in fast periodic visual stimulation streams are immune to temporal predictability

    Neuropsychologia 104:182–200.

    https://doi.org/10.1016/j.neuropsychologia.2017.08.010

    • PubMed
    • Google Scholar
69. 1. Rangarajan V
    2. Hermes D
    3. Foster BL
    4. Weiner KS
    5. Jacques C
    6. Grill-Spector K
    7. Parvizi J

    (2014) Electrical stimulation of the left and right human fusiform gyrus causes different effects in conscious face perception

    The Journal of Neuroscience 34:12828–12836.

    https://doi.org/10.1523/JNEUROSCI.0527-14.2014

    • PubMed
    • Google Scholar
70. 1. Ray S
    2. Crone NE
    3. Niebur E
    4. Franaszczuk PJ
    5. Hsiao SS

    (2008) Neural correlates of high-gamma oscillations (60-200 hz) in macaque local field potentials and their potential implications in electrocorticography

    The Journal of Neuroscience 28:11526–11536.

    https://doi.org/10.1523/JNEUROSCI.2848-08.2008

    • PubMed
    • Google Scholar
71. 1. Ray S
    2. Maunsell JHR

    (2011) Different origins of gamma rhythm and high-gamma activity in macaque visual cortex

    PLOS Biology 9:e1000610.

    https://doi.org/10.1371/journal.pbio.1000610

    • PubMed
    • Google Scholar
72. 1. Regan D

    (1966) Some characteristics of average steady-state and transient responses evoked by modulated light

    Electroencephalography and Clinical Neurophysiology 20:238–248.

    https://doi.org/10.1016/0013-4694(66)90088-5

    • PubMed
    • Google Scholar
73. 1. Retter TL
    2. Rossion B

    (2016) Uncovering the neural magnitude and spatio-temporal dynamics of natural image categorization in a fast visual stream

    Neuropsychologia 91:9–28.

    https://doi.org/10.1016/j.neuropsychologia.2016.07.028

    • PubMed
    • Google Scholar
74. 1. Retter TL
    2. Jiang F
    3. Webster MA
    4. Rossion B

    (2020) All-or-none face categorization in the human brain

    NeuroImage 213:116685.

    https://doi.org/10.1016/j.neuroimage.2020.116685

    • PubMed
    • Google Scholar
75. 1. Rice GE
    2. Hoffman P
    3. Binney RJ
    4. Lambon Ralph MA

    (2018) Concrete versus abstract forms of social concept: an fmri comparison of knowledge about people versus social terms

    Philosophical Transactions of the Royal Society B 373:20170136.

    https://doi.org/10.1098/rstb.2017.0136

    • Google Scholar
76. 1. Rossion B
    2. Torfs K
    3. Jacques C
    4. Liu-Shuang J

    (2015) Fast periodic presentation of natural images reveals a robust face-selective electrophysiological response in the human brain

    Journal of Vision 15:1.

    https://doi.org/10.1167/15.1.18

    • PubMed
    • Google Scholar
77. 1. Rossion B
    2. Jacques C
    3. Jonas J

    (2018) Mapping face categorization in the human ventral occipitotemporal cortex with direct neural intracranial recordings

    Annals of the New York Academy of Sciences 1426:e596.

    https://doi.org/10.1111/nyas.13596

    • PubMed
    • Google Scholar
78. 1. Rossion B
    2. Lochy A

    (2022) Is human face recognition lateralized to the right hemisphere due to neural competition with left-lateralized visual word recognition? A critical review

    Brain Structure & Function 227:599–629.

    https://doi.org/10.1007/s00429-021-02370-0

    • PubMed
    • Google Scholar
79. 1. Salado AL
    2. Koessler L
    3. De Mijolla G
    4. Schmitt E
    5. Vignal JP
    6. Civit T
    7. Tyvaert L
    8. Jonas J
    9. Maillard LG
    10. Colnat-Coulbois S

    (2018) SEEG is A safe procedure for A comprehensive anatomic exploration of the insula: A retrospective study of 108 procedures representing 254 transopercular insular electrodes

    Operative Neurosurgery 14:1–8.

    https://doi.org/10.1093/ons/opx106

    • PubMed
    • Google Scholar
80. 1. Sanada T
    2. Kapeller C
    3. Jordan M
    4. Grünwald J
    5. Mitsuhashi T
    6. Ogawa H
    7. Anei R
    8. Guger C

    (2021) Multi-modal mapping of the face selective ventral temporal cortex-A group study with clinical implications for ECS, ecog, and fmri

    Frontiers in Human Neuroscience 15:616591.

    https://doi.org/10.3389/fnhum.2021.616591

    • PubMed
    • Google Scholar
81. 1. Schroeder CE
    2. Lakatos P

    (2009) Low-frequency neuronal oscillations as instruments of sensory selection

    Trends in Neurosciences 32:9–18.

    https://doi.org/10.1016/j.tins.2008.09.012

    • PubMed
    • Google Scholar
82. 1. Schrouff J
    2. Raccah O
    3. Baek S
    4. Rangarajan V
    5. Salehi S
    6. Mourão-Miranda J
    7. Helili Z
    8. Daitch AL
    9. Parvizi J

    (2020) Fast temporal dynamics and causal relevance of face processing in the human temporal cortex

    Nature Communications 11:656.

    https://doi.org/10.1038/s41467-020-14432-8

    • PubMed
    • Google Scholar
83. 1. Sergent J
    2. Ohta S
    3. MacDonald B

    (1992) Functional neuroanatomy of face and object processing: a positron emission tomography study

    Brain: A Journal of Neurology 115 Pt 1:15–36.

    https://doi.org/10.1093/brain/115.1.15

    • PubMed
    • Google Scholar
84. 1. Talairach J
    2. Bancaud J

    (1973) Stereotaxic approach to epilepsy: methodology of anatomo-functional stereotaxic investigations

    Progress in Neurological Surgery 5:297–354.

    https://doi.org/10.1159/000394343

    • Google Scholar
85. 1. Tukey JW

    (1962) The future of data analysis

    The Annals of Mathematical Statistics 33:1–67.

    https://doi.org/10.1214/aoms/1177704711

    • Google Scholar
86. 1. Turner EC
    2. Young NA
    3. Reed JL
    4. Collins CE
    5. Flaherty DK
    6. Gabi M
    7. Kaas JH

    (2016) Distributions of cells and neurons across the cortical sheet in old world macaques

    Brain, Behavior and Evolution 88:1–13.

    https://doi.org/10.1159/000446762

    • PubMed
    • Google Scholar
87. 1. Vidal JR
    2. Ossandón T
    3. Jerbi K
    4. Dalal SS
    5. Minotti L
    6. Ryvlin P
    7. Kahane P
    8. Lachaux JP

    (2010) Category-specific visual responses: an intracranial study comparing gamma, beta, alpha, and ERP response selectivity

    Frontiers in Human Neuroscience 4:195.

    https://doi.org/10.3389/fnhum.2010.00195

    • PubMed
    • Google Scholar
88. 1. Volfart A
    2. Yan X
    3. Maillard L
    4. Colnat-Coulbois S
    5. Hossu G
    6. Rossion B
    7. Jonas J

    (2022) Intracerebral electrical stimulation of the right anterior fusiform gyrus impairs human face identity recognition

    NeuroImage 250:118932.

    https://doi.org/10.1016/j.neuroimage.2022.118932

    • PubMed
    • Google Scholar
89. 1. Wang Y
    2. Metoki A
    3. Smith DV
    4. Medaglia JD
    5. Zang Y
    6. Benear S
    7. Popal H
    8. Lin Y
    9. Olson IR

    (2020) Multimodal mapping of the face connectome

    Nature Human Behaviour 4:397–411.

    https://doi.org/10.1038/s41562-019-0811-3

    • PubMed
    • Google Scholar
90. 1. Winawer J
    2. Horiguchi H
    3. Sayres RA
    4. Amano K
    5. Wandell BA

    (2010) Mapping hv4 and ventral occipital cortex: the venous eclipse

    Journal of Vision 10:1.

    https://doi.org/10.1167/10.5.1

    • PubMed
    • Google Scholar
91. 1. Winawer J
    2. Kay KN
    3. Foster BL
    4. Rauschecker AM
    5. Parvizi J
    6. Wandell BA

    (2013) Asynchronous broadband signals are the principal source of the BOLD response in human visual cortex

    Current Biology 23:1145–1153.

    https://doi.org/10.1016/j.cub.2013.05.001

    • PubMed
    • Google Scholar
92. 1. Winawer J
    2. Parvizi J

    (2016) Linking electrical stimulation of human primary visual cortex, size of affected cortical area, neuronal responses, and subjective experience

    Neuron 92:1213–1219.

    https://doi.org/10.1016/j.neuron.2016.11.008

    • PubMed
    • Google Scholar
93. 1. Xing D
    2. Yeh CI
    3. Shapley RM

    (2009) Spatial spread of the local field potential and its laminar variation in visual cortex

    The Journal of Neuroscience 29:11540–11549.

    https://doi.org/10.1523/JNEUROSCI.2573-09.2009

    • PubMed
    • Google Scholar
94. 1. Yoshor D
    2. Bosking WH
    3. Ghose GM
    4. Maunsell JHR

    (2007) Receptive fields in human visual cortex mapped with surface electrodes

    Cerebral Cortex 17:2293–2302.

    https://doi.org/10.1093/cercor/bhl138

    • PubMed
    • Google Scholar

Author details

1. Corentin Jacques

   1. Université de Lorraine, CNRS, CRAN, Nancy, France
   2. Psychological Sciences Research Institute (IPSY), Université Catholique de Louvain (UCLouvain), Louvain-la-Neuve, Belgium

   Contribution

   Conceptualization, Software, Formal analysis, Investigation, Visualization, Methodology, Writing – original draft, Writing – review and editing

   Contributed equally with

   Jacques Jonas

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0001-8917-4346

2. Jacques Jonas

   1. Université de Lorraine, CNRS, CRAN, Nancy, France
   2. Université de Lorraine, CHRU-Nancy, Service de Neurologie, Nancy, France

   Contribution

   Conceptualization, Resources, Formal analysis, Investigation, Visualization, Methodology, Writing – original draft, Writing – review and editing

   Contributed equally with

   Corentin Jacques

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-1053-0748

3. Sophie Colnat-Coulbois

   Université de Lorraine, CHRU-Nancy, Service de Neurochirurgie, Nancy, France

   Contribution

   Resources, Investigation, Project administration

   Competing interests

   No competing interests declared

4. Louis Maillard

   1. Université de Lorraine, CNRS, CRAN, Nancy, France
   2. Université de Lorraine, CHRU-Nancy, Service de Neurologie, Nancy, France

   Contribution

   Resources, Funding acquisition, Project administration

   Competing interests

   No competing interests declared

5. Bruno Rossion

   1. Université de Lorraine, CNRS, CRAN, Nancy, France
   2. Université de Lorraine, CHRU-Nancy, Service de Neurologie, Nancy, France

   Contribution

   Conceptualization, Resources, Supervision, Funding acquisition, Investigation, Methodology, Writing – original draft, Project administration, Writing – review and editing

   For correspondence

   bruno.rossion@univ-lorraine.fr

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-1845-3935

• 935

  views

• 168

  downloads

• 4

  citations

Views, downloads and citations are aggregated across all versions of this paper published by eLife.