Article Figures & Data
Figures
- Figure 1.
Task description and behavioral performance. A, MSIT task description and timeline. A fixation cross appears on the screen 500 ms before stimuli are shown. Four labeled examples of the different types of interference are shown above the stimulus period in the task timeline. Feedback occurs following the subject's response to the stimulus. In valenced feedback trials, the target number turns green for correct responses and red for incorrect responses. In neutral feedback trials, the target number turns blue regardless of whether the subject responded correctly. Examples of the four possible types of feedback are shown for the example stimulus that contains both types of interference. B, Distributions of RTs for each subject colored by subject. C, Mean RTs across subjects for trials containing the four different types of interference. The asterisk indicates statistical significance over conditions for the ANOVA. D, Mean RTs across subjects for trials following correct and neutral feedback. n.s, No significant difference.
- Figure 2.
Depth electrode locations in seven subjects. The locations of the medial and lateral-most contacts on each depth electrode projected onto the medial and lateral surfaces of each hemisphere of the MNI standard brain. Colors represent individual subjects.
- Figure 3.
FRN in high- and low-frequency LFP. A, Response aligned spectrograms averaged over trials (mean ± SD = 289 ± 65.32) and patients. Frequency bands during the outcome period that are significantly different between feedback conditions are outlined with gray boxes (paired t tests, corrected for 72 hypotheses, all p < 0.01). B, Response aligned intertrial phase coherence averaged over trials and patients. C, Representative event related potentials from one subject for medial and lateral electrode contacts. Green and blue lines represent correct and neutral feedback conditions, respectively, and black lines represent the difference. Shading represents SE. D, Representative evoked high gamma response from one subject. E, Quantification of FRN amplitude from ERP (Mann–Whitney U test, p = 0.0041, N = 32 contacts). F, Quantification of high gamma power for each feedback condition (Mann–Whitney U test, p = 0.00096, N = 22 contacts). G, Quantification of latency differences between medial and lateral high gamma responses (two-sample t test, p = 0.02, N = 22 contacts). Asterisks indicate significant differences (p < 0.05). H, Lagged cross-correlogram between medial and lateral high gamma responses. The peak indicated that mean high gamma responses in lateral contacts lagged those in medial contacts by 79 ms.
- Figure 4.
PAC of low- and high-frequency LFPs. A, B, Representative PAC for a medial (A) and lateral (B) contact in one subject for each feedback condition (correct and neutral). C, D, Plots of the real (x-axis) and imaginary (y-axis) components of the modulation indices over trials for a representative patient for medial (C) and lateral (D) contacts. Blue dots represent the modulation indices for neutral trials and green dots represent the modulation indices for correct trials. E, Mean modulation indices across patients for medial and lateral contacts. Green and blue bars represent correct and neutral feedback trials, respectively. Asterisks indicate significant differences (p < 0.01). Modulation indices were greater for correct feedback trials than neutral feedback trials in medial (Tukey-Kramer, p = 0.0052, N = 36 contacts), but not in lateral (Tukey–Kramer, p = 0.0598, N = 36 contacts) contacts. The difference in theta-high gamma coupling between correct and neutral trials was greater in medial contacts than lateral contacts (Tukey–Kramer, p = 0.0098, N = 36 contacts).
- Figure 5.
Pairwise conditional GC between dACC and dlPFC. A, Pairwise GC over frequencies for the lateral-to-medial direction. Mean GC values across patients are shown in dark bold orange, and patient-specific data are shown in light orange. Peak frequencies averaged over patients are indicated in the colored text in the plot. B, Pairwise GC over frequencies for the medial-to-lateral direction. Mean GC values across patients are shown in dark bold blue, and patient-specific data are shown in light blue. Peak frequencies averaged over patients are indicated in the colored text in the plot.
- Figure 6.
CMI exchange between dACC and dlPFC. A, Mean CMI across subjects after feedback. Pink lines represent medial-to-lateral information transfer, and brown lines represent lateral-to-medial transfer. Gray lines represent shuffled data. Dashed and dotted lines show sinusoidal fits for medial-to-lateral and lateral-to-medial CMI, respectively. Vertical line represents feedback onset. B, Mean spectra of the CMI time series show increased theta power for medial-to-lateral CMI.
