Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2021 Feb 26;8(1):ENEURO.0406-20.2021.
doi: 10.1523/ENEURO.0406-20.2021. Print 2021 Jan-Feb.

Mapping Large-Scale Networks Associated with Action, Behavioral Inhibition and Impulsivity

L Fakhraei  1   2 , M Francoeur  1   2 , P Balasubramani  2 , T Tang  1   2 , S Hulyalkar  1   2 , N Buscher  1   2 , C Claros  1   2 , A Terry  1   2 , A Gupta  1   2 , H Xiong  2 , Z Xu  2 , J Mishra  2 , D S Ramanathan  3   2
Affiliations

Mapping Large-Scale Networks Associated with Action, Behavioral Inhibition and Impulsivity

L Fakhraei et al. eNeuro. .

Abstract

A key aspect of behavioral inhibition is the ability to wait before acting. Failures in this form of inhibition result in impulsivity and are commonly observed in various neuropsychiatric disorders. Prior evidence has implicated medial frontal cortex, motor cortex, orbitofrontal cortex (OFC), and ventral striatum in various aspects of inhibition. Here, using distributed recordings of brain activity [with local-field potentials (LFPs)] in rodents, we identified oscillatory patterns of activity linked with action and inhibition. Low-frequency (δ) activity within motor and premotor circuits was observed in two distinct networks, the first involved in cued, sensory-based responses and the second more generally in both cued and delayed actions. By contrast, θ activity within prefrontal and premotor regions (medial frontal cortex, OFC, ventral striatum, and premotor cortex) was linked with inhibition. Connectivity at θ frequencies was observed within this network of brain regions. Interestingly, greater connectivity between primary motor cortex (M1) and other motor regions was linked with greater impulsivity, whereas greater connectivity between M1 and inhibitory brain regions (OFC, ventral striatum) was linked with improved inhibition and diminished impulsivity. We observed similar patterns of activity on a parallel task in humans: low-frequency activity in sensorimotor cortex linked with action, θ activity in OFC/ventral prefrontal cortex (PFC) linked with inhibition. Thus, we show that δ and θ oscillations form distinct large-scale networks associated with action and inhibition, respectively.

Keywords: behavioral inhibition; brain mapping; impulsivity; local field potentials; orbitofrontal cortex; oscillations.

PubMed Disclaimer

Figures

Figure 1.
Figure 1.
Target electrode locations. Eight cannula each housing four microwires were implanted into the brain at different A/P and M/L locations. Each wire was measured and precut to reach a unique D/V location. This configuration provides 32 target locations within one hemisphere of the brain to record LFPs. All coordinates are calculated and shown relative to bregma (Paxinos and Watson, 2006). Each cannula (1–8) and each electrode wire (A–D) are shown on coronal rat brain sections modified from Paxinos and Watson (2006). Target sites are color coded by depth for reference (A = blue; B = green; C = yellow; D = red). The table contains AP, ML, DV and target location names for all 32 electrodes.
Figure 2.
Figure 2.
Task design and electrophysiological approach. A, The behavioral paradigm we used consisted of a visual stimuli instructing animals to respond immediately (go trial, denoted with an upward-facing striped rocket) or after specified delay (wait trial, denoted with a horizontal-facing white rocket). On go trials, animals had to respond within 2 s to earn a reward (20 μl of water). On wait trials, animals had to wait 2 s and then respond to collect a reward. B, A violin plot representation of the RT for go and wait trials across behavioral sessions (single dots represent individual behavioral sessions across 60 sessions). Animals could distinguish visual stimuli and showed appropriate behavior (waiting longer on wait compared with go trials). C, Histogram of all trials (from all sessions), demonstrating the difference in reaction time for wait and go trials, showing clear discrimination and attempts to wait. D, Average event-related potentials (ERPs) from a few selected brain regions: V1, M1, PPCx, and dmPFC from go and wait trials (n = 60 sessions). Data were broad-band filtered (0.5–500 Hz), and baseline normalized before averaging. Shaded-error plots show mean/SEM. There are significant differences between ERPs for go (blue) and wait (red) trials; *significant differences at p < 0.05, as estimated using a paired t test, followed by FDR-correction performed across the entire 2D time and electrode matrix. Extended Data Figure 2-1 shows preprocessing steps, number of trials/session, and a different view of electrode placement along histology.
Figure 3.
Figure 3.
Low-frequency activity linked with sensory-evoked responses. Activity in three key sensory-response mapping regions was investigated. A–D, TF activity from four motor regions was plotted: M1, parts of the FOF (A24b, A33), and anterior portion of M2. For each region we estimated average activity for correct go trials and the difference from correct go and correct wait trials (p < 0.05, FDR-corrected across all electrodes, times and frequencies). The third column shows the significant [go–wait] contrast at TF points that are also significant on go trials alone (statistical thresholded maps, FDR-corrected for each analysis separately before thresholding). Finally, for each region, we also plotted the mean/SEM trace within the δ frequency range for go (green) and wait (red) trials (time points with significant differences between the two denoted with *). E, Mean δ frequency activity across brain regions for go correct trials, and the statistically thresholded contrast [go–wait] map (thresholded for significant time points in the contrast that are also significant on go trials alone). Significant activations were observed in sensorimotor regions (ALM, M1, A33, A24a, A24b), along with thalamus, visual cortex, sensorimotor cortex, and NAcS. F, Mean/SEM of [go–wait] contrast averaged across δ frequencies from 300 to 500 ms poststimulus. All regions displayed are highly significant (adjusted p < 0.05, Bonferroni-correction for 32 electrodes). Extended Data Figure 3-1 includes extended mean/SEM and adjusted p values for all electrodes, and Extended Data Figure 3-2 shows results when calculated at the level of animals. G, Single trial example of δ activity from M1 to demonstrate what task-evoked δ oscillations looks like.
Figure 4.
Figure 4.
Low-frequency wPLI associated with action. The difference between go and wait trials in wPLI was calculated at δ frequencies, averaged over the period between 300 and 500 ms (analysis restricted to brain regions showing significant δ activity during this time period). A, We found increased connectivity (wPLI) in two distinct subnetworks at δ frequencies during this time period for go compared with wait trials. The first was a distinct motor network comprising of M1 (ALM) and a dorsal part of the FOFs (A24b). The second network comprised of sensory, reward and memory brain regions. Graph of the significant connections on the right. B, Bar-plots showing [go–wait] wPLI contrast between 300 and 500 ms poststimulus; *significant regions (FDR-correction applied across 11 × 11 matrix). Extended Data Figure 4-1 includes extended mean/SEM and adjusted p values for all electrodes. All bars show mean/SEM.
Figure 5.
Figure 5.
Response locked low-frequency activity differentiates motor brain regions. A, Low-frequency activity from go and wait trials, time-locked to the response, is shown for two motor brain regions: M1 and A33. Average activity shown from each session, contrasting the go and wait responses. M1 showed qualitatively similar low-frequency activations preceding the response. A33 shows generally stronger low-frequency activity preceding the responses for go compared with wait trials. B, Mean/SEM traces of the low-frequency activity for five motor regions involved in action displayed (M1, ALM, A24b, A24a, A33). A24a and A33 showed clearly larger differences in low-frequency activity for the delayed compared with immediate action trials. Significance marked, p < 0.05, FDR-corrected across time points.
Figure 6.
Figure 6.
θ Oscillations linked with behavioral inhibition. A, B, TF activity was plotted for two brain regions previously linked with inhibition: A32D and vOFC. Mean activity from wait-trials (left panel) and from the wait–go contrast was followed by the statistical thresholded [wait–go] contrast map (thresholded for significant activations for go trials alone and a positive difference for the wait–go contrast, both FDR-corrected). The statistically thresholded map shows greater θ activity for wait compared with go trial types (FDR-corrected, p < 0.05, with non-significant points set to 0). Line plots of the mean/SEM activity in the θ frequency band was displayed for go and wait trials, showing time points where the two regions have significantly different θ activity for wait compared with go trials. C, Whole-brain map of θ related to waiting, showing both mean wait-trial θ activity, and the significant, thresholded maps for wait–go (thresholded for significant θ activity for both wait trials and for the wait–go difference). D, Mean/SEM of the θ activity averaged across the 500- to 2000-ms window. All brain regions shown have significant activity for both wait trials and for the [wait–go] contrast (p < 0.05, FWE-corrected across 32 electrodes). E, Example trace of theta-filtered activity from OFC. Extended Data Figure 6-1 includes extended mean/SEM and adjusted p values for all electrodes, and Extended Data Figure 6-2 shows results when calculated at the level of animals.
Figure 7.
Figure 7.
wPLI associated with behavioral inhibition. A, We calculated the mean wPLI for each region to each other region for wait trials (thresholded using FDR-correction) and the [wait–go] difference (showing both the average wPLI and the FDR-corrected map, p < 0.05, with non-significant values set to 0). Extended Data Figure 7-1 plot the mean/SEM and p values for all electrodes. B, Graph of the wait–go difference shown at two different network thresholds to illuminate the strongest pair-wise wPLI connections (all shown are significant). Extended Data Figure 7-2 plot the mean/SEM and p values for all electrodes.
Figure 8.
Figure 8.
Correlation between network connectivity and successful inhibition. A, Logistic regression was performed over time between θ activity (averaged for successful vs unsuccessful wait trials) and trial-type from each of the behavioral sessions (FDR-corrected/thresholded across all electrodes/time). We plotted activity in 11 selected regions showing strong θ activity or connectivity. We found significant activity in DMS and M1 correlated with successful inhibition but relatively weakly. B, We performed a similar analysis using wPLI values averaged in an early period (between 500 and 900 ms after stimulus onset, FDR-corrected). We found highly significant and strong correlations with behavior using pair-wise connectivity. M1-OFC connectivity shows a strong and significant relationship with behavior. C, Multivariate regression model was developed with connectivity from M1 to the 10 other brain regions (only significant β values in the model shown), demonstrating that connectivity with ventral striatum and OFC are associated with improved impulsivity while connectivity with motor regions are associated with diminished impulsivity. Overall, multivariate regression model shows a strong relationship with behavior, demonstrating that connectivity can be used to accurately classify trial types. D, We next used an SVM ML model to measure classification of trial types using M1 connectivity at each time point for both θ and β activity. Model used a 75%/25% split (training vs test) with 10× cross-validation, and we randomized the initial 75/25 split ten times to produce a mean/SEM of the model. We performed this for wPLI from M1 for both theta (black) and Beta (orange) values. We found that the θ model achieves >80% performance by 500 ms poststimulus. We plotted the SVM Factor importance for the ML model (using time-points between 500-900ms post stim). We found connectivity values from vOFC, nAcC, VMS, DMS and A32D were particularly important in the prediction model.
Figure 9.
Figure 9.
Go wait task in humans. A, The human task was designed based on the rodent task, although with a more complicated set of stimuli. Go trials were indicated by a blue rocket, while any other color indicated a wait stimulus. Humans performed similarly to rats on go trials but were much better on wait trials. B, Action-related activity (contrast of [go–wait]) could be observed at δ frequencies in centro-parietal electrodes (average 200–300 ms poststimulation). The outline of the head in the scalp topography maps is not to scale and only shows head orientation. This activity source localizes to precuneus and postcentral gyrus. wPLI showed significant effects for this contrast primarily between centro-parietal electrode locations (p < 0.05). C, Activity linked with inhibition [wait–go] could be seen in frontal electrodes, particularly in θ frequencies (average 350–650 ms poststimulation). θ Activity source-localized to ventral and OFC areas, similar to that observed in rodents. Significant wPLI activity at θ frequencies could be observed between frontal electrode sites (p < 0.05).
See this image and copyright information in PMC

Similar articles

See all similar articles

Cited by

References

    1. Abela AR, Dougherty SD, Fagen ED, Hill CJR, Chudasama Y (2013) Inhibitory control deficits in rats with ventral hippocampal lesions. Cereb Cortex 23:1396–1409. 10.1093/cercor/bhs121 - DOI - PubMed
    1. Adhikari A, Topiwala MA, Gordon JA (2010) Synchronized activity between the ventral hippocampus and the medial prefrontal cortex during anxiety. Neuron 65:257–269. 10.1016/j.neuron.2009.12.002 - DOI - PMC - PubMed
    1. Balasubramani PP, Ojeda A, Grennan G, Maric V, Le H, Alim F, Zafar-Khan M, Diaz-Delgado J, Silveira S, Ramanathan D, Mishra J (2020) Mapping Cognitive Brain Functions at Scale. NeuroImage, in press. - PMC - PubMed
    1. Bari A, Robbins TW (2013) Noradrenergic versus dopaminergic modulation of impulsivity, attention and monitoring behaviour in rats performing the stop-signal task. Psychopharmacology (Berl) 230:89–111. 10.1007/s00213-013-3141-6 - DOI - PMC - PubMed
    1. Bari A, Dalley JW, Robbins TW (2008) The application of the 5-choice serial reaction time task for the assessment of visual attentional processes and impulse control in rats. Nat Protoc 3:759–767. 10.1038/nprot.2008.41 - DOI - PubMed

LinkOut - more resources