Focal gamma activity in frontal control regions revealed using intraoperative electrocorticography

How the frontal cortex is anatomically and functionally organized to control cognition remains puzzling. Numerous non-invasive brain imaging studies relate cognitive control to a localized set of frontal regions, part of a wider fronto-parietal network (FPN), that show increases in functional MRI (fMRI) signal during the performance of multiple cognitively demanding tasks. Lesions implicating frontal control regions lead to disorganized behaviour. However, the fMRI BOLD signal is an indirect measure of neuronal activity and represents evidence from a single modality. This has led to limited clinical translation of fMRI findings e.g. to guide the surgical resection of brain tumours. Here we sought supporting evidence for lateral frontal control regions using electrocorticography (ECoG). We recorded electrophysiological activity from electrodes placed on the lateral frontal cortex in patients undergoing awake craniotomy for glioma resection. During surgery, patients performed two verbal executive-related counting tasks with a difficulty level manipulation, closely adapting difficulty manipulations in fMRI studies of cognitive control. We performed spectral analysis focusing on the gamma range (30-250 Hz) due to mounting evidence of its value as an index of local cortical processing. Comparing hard versus easy demands revealed circumscribed frontal regions with power increases in the gamma range. This contrasted with spatially distributed power decreases in the beta range (12-30 Hz). Further, electrodes showing significant gamma power increases were more likely to occur within a canonical fMRI-defined FPN and showed stronger gamma power increases compared to electrodes outside the FPN, even at the single patient level. Reinforcing the need for careful task manipulation, an easy versus baseline comparison, which includes factors such as speech output, produced gamma changes over a wider area. Thus, using similar task difficulty manipulations, ECoG and fMRI signals converged on delineating lateral frontal control regions. These findings open the door for extending clinical functional mapping to the domain of cognitive control during awake neurosurgery.

Introduction regions that support cognitive control intraoperatively could provide valuable information to guide resection and prevent commonly-seen deficits following surgery (van Loon et al., 2015). In ECoG data recorded during surgery, we sought evidence for specific FPN activity linked to cognitive control demand.

Compared
to electroencephalogram (EEG) and magnetoencephalogram (MEG), ECoG offers both much-improved spatial resolution and the ability to detect local field potential (LFP) modulations within the broadband gamma frequency range . Accumulating evidence has emphasized the role of broadband gamma activity as an index for local cortical processing across a wide range of functional domains, but activity in this frequency range cannot usually be detected noninvasively because of its low amplitude and the low-pass filter induced by the scalp (Crone et al., 2006;Lachaux et al., 2012). In the cognitive control domain, evidence for the role of broadband gamma activity comes from invasive studies in both humans and animals. Increases in broadband gamma have long been known to reflect a heightened attentive state (Bouyer et al., 1981) and have been linked to processing of attended stimuli in both early sensory and fronto-parietal regions (Fries, 2001;Ray et al., 2008;Szczepanski et al., 2014;Helfrich and Knight, 2016). Increases in gamma power were also associated with working memory load, and during delay periods (Howard, 2003;Mainy et al., 2007), and sustained gamma responses were observed in frontal regions (Haller et al., 2018).
Responses to more abstract rules have been associated with frontal gamma increases (Voytek et al., 2015). A common limitation for the above-mentioned studies is their lack of anatomical specificity when localizing gamma activity, mostly allocating it to broad labels such as "dorsolateral prefrontal cortex". Such association cortices are rich in functionally heterogeneous regions (Glasser et al., 2016), making it difficult to relate electrophysiological and functional neuroimaging findings.
Increases in power in the gamma range are often accompanied by decreases in power in lower frequencies, in particular in the beta range (12)(13)(14)(15)(16)(17)(18)(19)(20)(21)(22)(23)(24)(25)(26)(27)(28)(29)(30). This has been demonstrated using different tasks and across many cortical regions including motor, sensory, visual and auditory tasks (Crone et al., 2006;Hermes et al., 2012b;Podvalny et al., 2015). For example, in motor and sensory cortex, hand movements were associated with both increases in high-frequency band power as well as decreases in the beta range (Hermes et al., 2012b). More recently, increases in the gamma range and decreases in the beta range in the prefrontal cortex of non-human primates have been associated with working memory, a central component of cognitive control processes (Lundqvist et al., 2016).
To identify an electrophysiological signature of domain-general control regions in the human brain, we used ECoG during awake brain surgeries while patients performed two tasks with a varying level of control demand. Following a previous demonstration of the feasibility of this approach (Erez et al., 2021), here we use dense coverage of the frontal cortex to characterize the power modulations associated with increased cognitive demand, and directly link them to the FPN.
Previous studies demonstrated correlates between the fMRI BOLD response and broadband gamma power (Logothetis et al., 2001;Mukamel et al., 2005;Siero et al., 2013). Taken together with reports of power changes across a variety of tasks related to cognitive control processes (Crone et al., 2006;Lachaux et al., 2012), we predicted that increased control demand will be associated with power increases in the broadband gamma range, especially within the FPN, accompanied by power decreases in the beta range.
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted August 13, 2021. ;https://doi.org/10.1101https://doi.org/10. /2021 Methods Patient recruitment Twenty-one patients were recruited from the pool of glioma patients who have been indicated to undergo awake resections of their tumours at the Department of Neurosurgery at [redacted to comply with preprint server policy]. Data from thirteen patients were complete and were included in the study (age range 22-56; 6 males; see Table 1 for patient demographics). Data from the remaining eight patients were excluded either due to technical difficulties (n=6) or an inability to perform the tasks during the surgery (n=2). All study procedures were approved by the East of England -Cambridge Central Research Ethics Committee (REC reference 16/EE/0151). All patients gave written informed consent to participate and were aware that the research would not benefit themselves, or impact their clinical care before, during or after surgery.

Experimental procedures
The patients were familiarized with the tasks during standard pre-operative clinic visits and as part of a pre-operative research-dedicated assessment. During the surgery and following the craniotomy, testing was performed after the patient had been awakened and prior to tumour resection. In one patient, the testing was performed after partial resection due to clinical considerations. Figures 1a,b illustrate the intraoperative setup and cognitive tasks. During testing, all personnel in the surgical theatre were asked to limit their conversations to minimize disruptions. Patients performed one baseline task and two cognitive tasks. For the baseline task, the patients were asked to stay calm and remain silent for a period of 2-3 minutes (rest). The two cognitive tasks were simple counting (1 to 20; easy) and alternate counting and reciting the alphabet (1, a, 2, b, 3, c, up to 20; hard). Each task condition was repeated for 2-5 trials (median for both = 4 trials) based on each patient's ability and time constraints during the surgery. Trial onset and offset markers were manually recorded on the acquisition system. Trial durations were 20.1±7.4s and 29.4±9.4s for the easy and hard conditions, respectively. Most . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted August 13, 2021. patients were instructed to alternate between trials of the easy and hard conditions, though on a few occasions some easy/hard trials were performed in succession.
Only correctly performed trials (i.e. no errors in simple or alternate counting) were included in the analysis (e.g. a failed hard trial that was excluded: 1, a, 2, b, 3, b, 4, b, 5, b, 6, b…).

Electrode localization
The extent of craniotomy of all patients was determined by clinical considerations to allow for tumour resection. Based on the craniotomy size and location, one to three electrode strips were placed on the cortical surface in regions judged by the neurosurgeon to be healthy (i.e. macroscopically not containing tumour). Strips placed on the tumour or outside of the frontal and motor cortices were excluded from analysis. Each strip was composed of four electrodes. Two types of strips were used with electrode diameter either 5 mm (MS04R-IP10X-0JH, Ad-Tech, Medical Instruments corporation, WI, USA) or 3 mm (CORTAC 2111-04-081, PMT Corporation, MN, USA). For both strip types, electrode spacing was 10 mm centre to centre.
Electrode locations were determined either using (1) an automated method with a probe linked to a stereotactic neuronavigation system (StealthStation® S7® System, Medtronic, Inc, 24 Louisville, CO, USA) or (2) a semi-manual grid method using intraoperative photographs and a grid-like delineation of cortical sulci and gyri. Most electrodes (51/79) were localized using the automated method, and due to occasional technical limitations, 28 electrodes were localized using the grid method. Both methods are detailed below.
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted August 13, 2021. ;https://doi.org/10.1101https://doi.org/10. /2021 (1) Stereotactic neuronavigation: A hand-held probe was placed at the centre of each electrode, automatically registering its physical coordinates, using the neuronavigation system, to the patient's native high resolution preoperative T1w scan. In some cases, coordinate data were available for only two or three out of the four electrodes in each strip. This was due to either time constraints during the surgery or because an electrode was located underneath the skull, excluding probe placement. Each patient's native T1w scan was linearly co-registered with the MNI template volume at 2 mm resolution using FLIRT as implemented in FSL (Jenkinson et al., 2012) using 12 degrees of freedom (full set affine transformation) and the correlation ratio cost function. The resulting native-to-MNI transformation matrix was then used to convert native electrode coordinates to MNI coordinates.
(2) The grid method: This follows the method described in (Havas et al., 2015) and (Ojemann et al., 1989). (a) Visible major sulci were delineated on the intraoperative photographs: precentral sulcus, sylvian fissure, inferior and superior frontal sulci. Spaces between these sulci were populated by vertical lines (1.5 cm apart) to create a grid-like structure. (b) A grid was created in the same way on a template cortical reconstruction of the MNI volumetric map (reconstructed using the Electrode displacements due to brain shifts following the craniotomy were compensated for by back-projecting the electrode locations onto the cortical surface along the local norm vector (Hermes et al., 2010) as implemented in the fieldtrip (v20160629) protocol for human intracranial data (Stolk et al., 2018).

Electrophysiological data acquisition and analysis
Data were recorded using a 32-channel amplifier (Medtronic Xomed, Jacksonville, FS, USA) sampled at 10 KHz. Potential sources of electrical noise such as the . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted August 13, 2021. ; https://doi.org/10.1101/2021.08.13.21261980 doi: medRxiv preprint surgical microscope, patient warming blanket, and IV pumps were identified and repositioned to avoid signal contamination. The data were recorded via dedicated channels on the acquisition system and two Butterworth online filters were applied: a high-pass filter at 1 Hz and a low-pass filter at 1500 Hz. A ground needle electrode was connected to the deltoid muscle and the electrodes were referenced to a mid-frontal (Fz) spiral scalp EEG electrode.
Data were analysed offline using EEGLAB (v13.6.5b) and custom MATLAB scripts. The data were downsampled to 2 kHz then re-referenced using a bipolar scheme to detect any activity changes with the highest spatial resolution as well as to avoid contamination of high frequency signals by scalp muscle artefacts detected by the Fz electrode. The last electrode on the strip was excluded from analysis; i.e.
for a four-electrode strip, electrode pairs 1-2, 2-3 and 3-4 were used and assigned to electrode positions 1, 2 and 3, respectively. The location of electrode 4 was discarded. Thus, out of the original 79 electrodes, re-referenced data from 59 were used for further analysis (Figure 1c). Out of these 59 electrodes, 41 were on the left hemisphere and 18 on the right. 25 electrodes were placed on the middle frontal gyrus (MFG), 28 on the inferior frontal gyrus (IFG) and on 6 on motor cortex.
A notch filter was applied at 50 Hz and its harmonics to remove line noise.
Notch filtering was also applied at 79 Hz and its harmonics to remove additional noise observed in the data, probably due to equipment in the surgical theatre. Data were then bandpass filtered into 6 classical frequency bands (delta: 1-4 Hz, theta: 4-8 Hz, alpha: 8-12 Hz, beta: 12-30 Hz, low gamma (LG): 30-70 Hz, high gamma (HG): 70-250 Hz). Instantaneous power of the timeseries was obtained by squaring the absolute amplitude envelope of the Hilbert transformed data.
The power timeseries data were then segmented into separate conditions and trials. Because trial onset and offset markers were manually recorded, 2s from the beginning and end of the rest trial and 1s from each task trial were excluded to account for human reaction time related error. For the hard trials (alternate numbers/letters), a further 3s from the beginning of each trial was excluded to discard the initial easy phase of this task (1, a, 2, b, 3, c,). One power value for each condition was obtained by concatenating its data across trials and averaging across time points. For each pair of conditions (hard>easy, easy>rest, hard>rest), the percentage signal change was computed as: [(power in condition 1/power in condition 2) -1] * 100.
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted August 13, 2021. ; https://doi.org/10.1101/2021.08.13.21261980 doi: medRxiv preprint For each electrode, a permutation testing approach was used to statistically compare power change across each pair of conditions. For each electrode, the instantaneous power timeseries of all task trials from both conditions were concatenated serially to form a loop in the same order in which they were conducted. To close the loop, the end of the last trial was joined to the beginning of the first trial. All trial onset/offset markers were then shifted using the same jitter (randomly generated for each permutation), allowing them to "rotate" along the data loop. This rotation approach was used to generate surrogate power data while preserving trial lengths and the temporal correlations in the data. After the rotation, we computed the mean power (for each condition) and power ratio (across conditions) based on the new trial markers. Applying this rotation approach on the timeseries of the power rather than the raw data ensured that there were no artefacts in the form of sudden power changes at the points of trials concatenation. This process was repeated 100,000 times to create a surrogate distribution against which two-tailed statistical significance could be calculated (percentile ranks 97.5 and 2.5) for each electrode.
To relate electrode locations with control-related regions identified by fMRI studies, a canonical resting-state fronto-parietal network (FPN) volumetric mask was used (Yeo et al., 2011). To match the original resolution of electrode localization using the grid method, the volumetric mask was resampled to 2 x 2 x 2 mm, then binarized to include any voxel with a non-zero probability. A volumetric approach was used rather than a surface-based approach because the presence of cortical lesions (i.e. the tumour) prevented accurate cortical surface extraction.
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted August 13, 2021. ; is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted August 13, 2021. ; Figure 1. Intraoperative ECoG setup and electrode localization. (a) Intraoperative setup: Patient is awake during the three experimental conditions and the electrophysiological signals are simultaneously recorded using electrode strips placed directly on the cortical surface. (b) Experimental conditions: one rest (no task) and two verbal counting tasks. Easy task involved simple counting from 1 to 20. Hard task involved alternating between numbers and letters. (c) Left: Electrode distribution for each patient in a separate colour (13 patients, 59 electrodes after bipolar re-referencing). Right: hemispheric distribution of electrodes.
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

Results
Hard>easy contrast reveals local increases in gamma power contrasting with widespread decreases in beta power First we examined spectral power changes for the hard>easy contrast. As discussed above, this contrast commonly shows frontal control regions in fMRI studies. Based on previous reports in the literature, we focused on power changes in the two gamma bands (HG and LG) and the beta band. For the HG and LG bands, statistically significant electrodes showed predominantly HG and LG increases (Figure 2a, 2b) and almost all electrodes showing significant LG increases also showed HG increases (92.3%) (Figure 2a). In contrast, almost all significant changes in the beta band were power decreases (Figure a, 2b). Most electrodes showing significant HG and LG increases also showed significant beta decreases (85.1% HG, 76.9% LG). While these results suggest that increases in gamma power co-occur with decreases in beta power, across all electrodes there were only modest correlations between gamma increases and beta decreases (for HG, r=0.64, p<0.0001; for LG, r=0.32, p=0.005).
Next we assessed the spatial distribution of these power modulations. Figure   2a, b shows circumscribed gamma power increases, while beta decreases are more broadly distributed across the recording area. To quantify this observation, we compared the Euclidean distances between all pairs of electrodes with significant power changes for each of the three bands (unpaired t-test, HG vs beta t 877 =-7.8, p=1.5x10 -14 , LG vs beta t 604 =-2.2, p=0.03; unpaired t-test after excluding electrodes lying on the motor cortex, HG vs beta t 784 =-3.5, p=0.00046, LG vs beta t 511 =-0.98, p=0.33). Although this measure is, at least in part, affected by the spatial distribution of the electrodes, it nevertheless captures the distributed nature of the beta decreases compared to the more focal increases in the gamma range.
For completeness, we also examined power modulations in the lower frequency bands (delta, theta, alpha). These bands showed a similar picture to the beta band, with predominantly power decreases though with a patchier spatial arrangement (Supplementary Figure 1). Overall, these results show that the . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
To demonstrate that our selected canonical frequency bands do not bias the results, Figure 2c shows the average power spectral density (PSD), for each condition, of all the electrodes that showed significant changes in all three bands: HG and LG increases as well as beta decreases. Indeed, the power modulations computed using canonical frequency bands align closely with observed differences in PSDs between hard and easy tasks. Furthermore, the lack of PSD peaks in the gamma range suggests that the simultaneous power increases in HG and LG reflect a broadband process (Miller et al., 2009).
Taken together, these results show that, along the lateral frontal cortex, increased cognitive demand is associated with a localized increase in high frequency power and a spatially distributed decrease in low frequency power.
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted August 13, 2021. ; . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

High frequency power increases overlap with the fMRIdefined fronto-parietal network
The results so far highlight a localized group of electrodes with increases in gamma power surrounded by dorsal and posterior electrodes that show little to no power modulations (Figure 2a). This spatial pattern is reminiscent of fMRI studies identifying localized frontal control-related patches. Here we predicted that gamma increases in our data would overlap with a canonical mask of the control-related FPN as defined in fMRI data (taken from Yeo et al., 2011b; see Methods).
In our dataset, 33 electrodes were located within the FPN, while 26 lay outside this network (Figure 3a). Of the FPN electrodes, 61.3% showed a significant HG increase in the hard>easy contrast, as compared to 28.6% of non-FPN electrodes (Figure 3b).
In contrast, the proportions of electrodes showing beta power decreases were similar for FPN and non-FPN (51.6% vs 60.7%). Therefore, while power increases with increased demand were more likely within the FPN compared to outside the network, decreases in the beta band were equally likely both within and outside the network.
In a subsequent analysis, we examined data across all electrodes, irrespective of individual significance. Here too, HG and LG power increases were significantly stronger in FPN than in non-FPN electrodes (Figure 3c; unpaired t-test HG t 57 =2.66, p=0.005, LG t 57 =2.71, p=0.004, p-values are one-tailed due to our directional prediction of gamma increases within the FPN). Again, there was no significant difference between FPN and non-FPN electrodes for beta power decreases (t 57 =1.2, p=0.2, two tailed as we had no specific directional prediction) nor for any of the lower frequency bands (two tailed p>0.6).
A limitation of the previous analysis is that most effects could be driven by a subset of patients. To address this, we used data from all the patients who each had electrodes overlapping with both FPN and non-FPN regions to perform a withinsubjects analysis (11 out of the 13 patients). Despite the lower number of patients available for this analysis, the results again showed that within the same patient, FPN electrodes showed statistically stronger LG power increases compared to non-FPN electrodes (Wilcoxon signed rank test; p=0.02, one tailed) and no differences . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
Collectively, these results suggest that demand-related power increases in high frequency bands co-localize with the fMRI-identified FPN. Thus, results from two different modalities (ECoG and fMRI) converge in defining a localized region within the lateral frontal lobe related to cognitive control.
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted August 13, 2021. ; . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted August 13, 2021. ;https://doi.org/10.1101https://doi.org/10. /2021 Power modulations in the easy>rest contrast Next we sought to assess whether the localized gamma increases, identified in the hard>easy contrast where the conditions are matched on multiple aspects (e.g. speech, counting), would also replicate in the easy>rest contrast where those aspects are not matched.
The distribution of HG, LG and beta changes in the easy>rest contrast is shown in Figure 4a. Across all electrodes, there was no significant correlation between gamma power increases for hard>easy and easy>rest (HG: r=-0.07 p=0.53; LG: r=-0.09 p=0.42), indicating different topographical arrangement for the gamma modulations in the two contrasts. For beta power decreases there was a weak correlation (r=0.22, p=0.045). In contrast to the results for the hard>easy contrast, the percentage of electrodes with significant gamma increases was similar between FPN and non-FPN electrodes (Figure 4b, HG 32.3% vs 39.3%; LG 22.6% vs 17.9%). Similarly, neither the average power modulations across electrodes (Figure   4c; unpaired t-test HG t 57 =-1.28, p=0.90, LG t 57 =-0.67, p=0.78, one-tailed; beta t 57 =-0.69, p=0.49 two tailed) nor the within-subject analysis (Figure 4d; Wilcoxon signed rank test, HG p=0.86, LG p=0.58, one tailed for both; beta p=0.52 two tailed,) showed any significant differences between FPN and non-FPN electrodes across the three bands. Again these results were stable after excluding the apparent outlier patient in Figure 4d (within-subject analysis: Wilcoxon signed rank test, HG p=0.99; LG p=0.75, one tailed for both; beta p=0.77 two tailed).
Interestingly, easy>rest gamma increases were shifted posteriorly to hard>easy increases (Figure 4e). To quantify this, we compared y-coordinates Altogether, these results confirm that the hard>easy contrast is most suitable for revealing localized control-related gamma frontal regions.
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted August 13, 2021. ;  Figure 3(b-d). (e) Box plots of y-coordinate of electrodes with significant power increases. Middle black line: mean; darker box limits: 1 SD; lighter box limits: 95% CI. *p<0.05 +p=0.07, using Wilcoxon rank sum test.
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

Discussion
We recorded LFP signals using ECoG from the lateral frontal surface of human patients undergoing awake craniotomies for tumour resections. The results revealed a circumscribed frontal region that shows increases in broadband gamma power with increased task demands. Localised gamma increases were accompanied by spatially broad beta power decreases. Electrodes showing gamma increases overlapped with a canonical mask of the fMRI-defined FPN. Our results confirm that specific control regions of the lateral frontal surface, previously defined by fMRI, can also be identified using clinically-relevant electrophysiological recording.
Delineating lateral frontal control regions using converging evidence from ECoG and fMRI There is some evidence from previous electrophysiology studies that gamma power increases in frontal regions are associated with multiple executive processes such as attention, working memory and abstract rules (Fries, 2001;Ray et al., 2008;Szczepanski et al., 2014;Helfrich and Knight, 2016). However, invasive human studies have rarely systematically assessed the anatomical distribution of gamma increases, with most studies reporting isolated findings in individual electrodes (Lachaux et al., 2012;Helfrich and Knight, 2016). This lack of detailed topographical assessment limits our ability to relate findings across studies and across neuroimaging modalities. Here we demonstrate that a circumscribed region along the lateral frontal surface robustly shows increases in gamma power during an executive task. Although some previous studies have reported different reliabilities and putative physiological origins for HG and LG (Crone et al., 2006;Ray and Maunsell, 2011;Buzsáki and Wang, 2012;Lachaux et al., 2012), in our data both gamma bands were broadly in agreement.
Our results suggest converging evidence from ECoG and fMRI modalities for control-related parcellation of the lateral frontal cortex. We used a task demand manipulation similar to the one that was previously shown to recruit control regions in fMRI studies (Fedorenko et al., 2013;Assem et al., 2020;Shashidhara et al., . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted August 13, 2021. ;https://doi.org/10.1101https://doi.org/10. /2021 2020) and the observed gamma power increases associated with increased demand co-localized with a canonical FPN derived from independent resting-state fMRI data from healthy participants. Previous studies have linked localized gamma increases with corresponding fMRI activations in early cortical regions (Nir et al., 2007;Engell et al., 2012;Hermes et al., 2012a). Accordingly, gamma increases were generally interpreted to reflect localized task-related processing. The current results extend these reports to the domain of cognitive control in frontal regions.
In contrast to focal increases in the gamma range, we observed spatially broad decreases in power in the beta band (and other lower frequencies) which were not confined to the FPN. It has been previously shown that synchronization in lower frequency bands between fronto-parietal regions improved during executive tasks e.g. (Voytek et al., 2015). Our findings suggest that increases in beta synchrony may be accompanied by decreases in power. Concerning the relation between gamma increases and beta decreases, one suggestion is that these reflect two sides of the same process, a rotated power spectrum around a middle range frequency (Podvalny et al., 2015;Helfrich and Knight, 2016). Recent evidence, however, argues against this simple interpretation, showing that depending on the cortical region, increases in high gamma power are not necessarily accompanied by decreases in low frequency power (Fellner et al., 2019). In line with this, the current results also showed that beta decreases were more spatially broad and were not necessarily accompanied by gamma increases. Another framework proposes a hybrid spiking-synaptic plasticity working memory model, in which bursts of spikes (gamma increases) in superficial layers serve to encode and maintain working memory content, while beta, which is assumed to have an inhibitory role, is suppressed in deeper layers to allow superficial gamma bursts (Lundqvist et al., 2011;Miller et al., 2018). This model is specific for regions that are involved in working memory processes, and our results are in line with this suggestion within the control-related areas. As it stands, however, the model does not predict decreased beta power in regions outside the FPN. Pending further experimental and theoretical studies, our results provide an important constraint for physiologicallymotivated models of executive processes.
Power modulations in the gamma range for the easy>rest contrast revealed posterior shift of gamma increases compared to the hard>easy contrast. One possibility is that this shift mirrors fMRI studies showing an anterior-posterior . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted August 13, 2021. ;https://doi.org/10.1101https://doi.org/10. /2021 cognitive demand gradient across the lateral frontal surface, with easier tasks involving more posterior regions and more challenging tasks recruiting more anterior regions (Badre and Nee, 2018;Shashidhara et al., 2019;Assem et al., 2020). Another possibility is related to the conditions that comprise the two contrasts. While the hard>easy contrast compared conditions that are closely matched on features such as speech and counting task, the easy>rest contrast is less controlled. Accordingly, changes in activity may be related to several differences between the conditions, most prominently speech production (e.g., the easy condition includes speech while the patient remains silent during the rest condition).
Plausibly, posterior gamma increases might also reflect articulation or language related processing in the posterior inferior frontal gyrus (Basilakos et al., 2018).

Clinical potential
Our findings open the door for extending clinical functional mapping to the domain of cognitive/executive control during tumour resection surgeries. Damage to control regions is associated with disorganized behaviour (Glascher et al., 2010, Woolgar et al., 2010Warren et al., 2014) and poorer recovery following neurosurgery (Romero-Garcia et al., 2019). During surgeries, electrical stimulation is commonly used to map motor and language functions, where activity is confined to well-defined areas with clear behavioural responses, so that resection of eloquent tissue can be avoided. In higher association areas and in particular for executive control regions, the use of stimulation is more challenging, primarily because of the complex nature of the mapped functions and the distributed areas that support them.
In addition, intraoperative stimulation is time-consuming while neurosurgeons search for the areas responsive to the tested functions with increased risk of seizures. Current mapping approaches to assess executive regions intraoperatively are limited, with only sparse evidence where direct electrical stimulation was used (Wager et al., 2013;Puglisi et al., 2018Puglisi et al., , 2019. ECoG has the potential to provide complementary information to guide stimulation and clinical decision making, especially for executive control functions. Most ECoG studies that investigate cognitive control are conducted at the patient's bedside between two surgeries as part of epilepsy treatment, and usually use complex computer-based tasks that are not suitable for an intraoperative setting. Here we employed a simple behavioural . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted August 13, 2021. ; https://doi.org/10.1101/2021.08.13.21261980 doi: medRxiv preprint manipulation based on task demand which was well tolerated by most patients and the whole testing was performed in under 10 minutes. The link of increased gamma power to the FPN establishes a critical anatomo-functional foundation for this approach. These results also suggest the FPN remains relatively intact next to tumour tissue. Further, the detected gamma signals are characterized by spatial specificity of a few millimetres. For example, nearby electrodes on the same strip can show differential gamma responses. This level of neuroanatomical precision is vital for guiding neurosurgeons during tumour excision.
Exploring several future avenues will help scale up this approach for clinical mapping. For example, while the current study aimed to detect an index of cognitive control processing, only one type of cognitive demand was manipulated. A conjunction of activity across several tasks is more likely to focus on core domaingeneral control regions (Assem et al., 2020). This will require modifying existing paradigms to more surgical theatre friendly versions. One such task could be contrasting verb generation with repeating nouns. Relatedly, our electrodes coverage was mostly limited to the middle and inferior frontal gyri. It will be clinically useful to extend the current findings to other components of the FPN (frontal pole, dorsal frontal and lateral parietal), for example in cases of different or distributed tumour locations. Similarly, a denser and more distributed set of electrodes would allow investigating electrophysiological signals in fMRI-based networks neighbouring the FPN. Further, only amplitude modulations were investigated in this study. Investigating synchrony within and across frequency bands (e.g., phase changes, phase-amplitude coupling) could reveal unique markers of FPN dynamics during task execution. Precise neurosurgical resection would also benefit from individually specific fMRI defined FPN (Shashidhara et al., 2020). This is complicated by our poor understanding of how brain tumours affect BOLD signal and shift the topographies of functional networks. Ultimately, integrating ECoG with brain stimulation and pre-operative multi-modal brain imaging opens up a new exciting phase for comprehensive surgical planning (Hart et al., 2020).
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted August 13, 2021. . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted August 13, 2021. ; https://doi.org/10.1101/2021.08.13.21261980 doi: medRxiv preprint

Supplementary Figures
Supplementary Figure 1. Hard>easy power modulations in lower frequency bands. Electrodes with significant (p<0.05, uncorrected) power increases (red), decreases (blue) and non-significant changes (white).
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted August 13, 2021. ; https://doi.org/10.1101/2021.08.13.21261980 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted August 13, 2021. ; https://doi.org/10.1101/2021.08.13.21261980 doi: medRxiv preprint