Does anterior cingulate cortex damage reduce the reward positivity in humans?

The RewP is a scalp signature of goal completion

The reward positivity is an event-related potential, an averaged voltage change in scalp EEG time-locked to a stimulus. It appears over frontocentral electrodes roughly 250 to 350 ms after feedback tells a person whether an ongoing event went better or worse than expected. The paper stresses that the name is narrower than the phenomenon: the RewP is produced by abstract performance feedback, not only by events with hedonic value, so the authors characterize it as an electrophysiological signature of goal completion.

The measurement itself is a difference wave. The researchers subtracted the ERP after positive feedback from the ERP after negative feedback, which removes activity common to both conditions and isolates the component tied to feedback valence. In this study the RewP was read at a virtual FCz channel built by averaging channels Fz and Cz, because the recording system had no electrode at FCz itself. Amplitude was taken as the most negative deflection of the difference wave within a 250–400 ms window, and because the RewP is expressed as a negative-going deflection, more negative values mean a larger RewP.

Scalp EEG alone cannot locate the generator

Whether the ACC generates the RewP has been disputed because of the inverse problem: an infinite number of neural source configurations can produce the same voltage pattern at a given scalp electrode. Earlier source-localization work on scalp EEG pointed to ACC, but other studies placed RewP sources in posterior cingulate cortex or the basal ganglia. fMRI, combined EEG/fMRI, and animal recordings have also implicated ACC, yet each method carries its own limits, including the coarse temporal resolution of the blood-oxygen-level-dependent signal and uncertain homologies between animal and human brains.

The authors' own prior intracranial EEG work offered more direct evidence. Pooling current source density data across epilepsy patients, they found that a circumscribed region of caudal ACC spanning the cingulate and paracingulate sulci, including Brodmann areas 24c' and 32', best accounted for the time course of the scalp RewP. That correlation still left the causal question open, which is what the lesion study was designed to address.

A lesion test of a single generator

The logic is a loss-of-function prediction: if caudal ACC generates the RewP, then damage there should attenuate it. Seventy-two stroke patients with focal frontal lesions were enrolled prospectively, and 68 (94.4%) completed the virtual T-maze guessing task with scalp EEG and entered analysis. Mean age was 60.4 years, 47.1% (n = 32) were female, and 17 patients (25%) had ACC lesions, split between 8 rostral and 9 caudal.

Lesion locations were mapped onto a standard brain template, and non-parametric voxel-based lesion-symptom mapping tested where damage was associated with smaller RewP amplitude, correcting for age and lesion volume. Only voxels damaged in at least 5% of the sample, meaning at least 4 of 68 patients, were included, and tests were one-tailed because the hypothesis was that injury would only shrink the RewP. The primary analysis restricted itself to five atlas regions spanning the caudal ACC area, an a priori choice justified by the earlier intracranial and fMRI findings.

Right caudal ACC damage shrinks the RewP

Across all 68 patients the grand average RewP peaked at 319 ms after feedback onset, with a peak amplitude of −1.1 µV and a frontocentral scalp distribution slightly lateralized to the right. Two right-hemisphere regions survived the permutation analysis: right frontal superior medial cortex (z = 1.96, P < 0.05) and right cingulate cortex 1 (z = 1.85, P < 0.05). Damage to these regions was significantly associated with a reduced RewP.

The group comparison makes the size of the effect concrete. Six patients whose lesions fell inside the right caudal ACC region of interest produced a grand average RewP peaking at −0.05 µV, while the 62 patients with lesions outside that region produced a normal-looking RewP peaking at −0.81 µV. The uncorrected voxelwise map pointed to the same area, with its highest z-score (2.4) at right caudal ACC, though that voxel did not survive correction for multiple comparisons.

No wider frontal-lobe effect appeared

Because the single-generator hypothesis is contested, the authors ran two exploratory analyses. A coarser atlas covering the whole medial wall, including caudal ACC, again flagged the right superior medial frontal cortex, consistent with the primary result. A final analysis spanning 10 frontal-lobe regions outside the midline produced no significant associations, even though the lesion overlap map showed damage distributed widely across frontal cortex. That negative result argues against the RewP being generated elsewhere in the frontal lobe, and the authors note it is unlikely to reflect a statistical bias in the raw data.

The finding is right-hemisphere only

The effect was limited to the right hemisphere, whereas the authors' earlier intracranial study was lateralized to the left. They propose that the left caudal ACC lacked sufficient statistical power: overlap density was higher in the right caudal ACC region (4 patients) than in the left (3 patients), which fell below the inclusion threshold. The authors explicitly label this interpretation post-hoc and call for further research. They also defend analyzing hemispheres separately, stating there is currently no evidence that left and right ACC function dependently.

Other design choices carry their own limits. Patients were recruited in both acute and chronic phases to maximize sample size, so RewP amplitude could have differed between phases in ways that affected results. For three participants with ambiguous waveforms or posterior scalp distributions, amplitude was set to zero, a decision informed by expert opinion but acknowledged as somewhat subjective; the researchers were unaware of lesion locations when computing amplitudes. The small number of patients with lesions inside the caudal ACC region, six in the illustrative comparison, is the central constraint on how precisely the effect can be localized.

What the causal claim rests on

The paper's causal language rests on convergence rather than on any single analysis. A prior intracranial current source density study identified caudal ACC as the best account of the scalp RewP time course; this study shows that damage to the same region, centered on the cingulate sulcus and extending to the paracingulate sulcus, attenuates the component. The secondary analyses rule out other frontal midline and lateral frontal locations within the resolution of the atlases used. Together these support a single generator in caudal ACC, which the authors connect to the region's proposed role in motivating extended behaviors.

The authors also raise a possibility their design cannot settle. In the group with caudal ACC damage, ERPs following the RewP were generally suppressed as well, hinting that the group difference may extend beyond the RewP itself. Because ACC is widely viewed as a hub of the salience network, the authors suggest that disrupting the RewP could affect the motivational salience of feedback passed downstream. They note that ACC damage could instead selectively affect working memory or attention required for feedback processing, though they consider this unlikely given current evidence that ACC lesions do not produce widespread behavioral impairments.

Where this changes how to read a RewP

For anyone interpreting RewP amplitude as an index of ACC function, the practical upshot is anatomical and lateral: in this sample, a blunted RewP tracked damage to right caudal ACC, specifically areas 24c' and 32', and not damage elsewhere in the frontal lobe. That makes the component more informative about a specific medial frontal hotspot than about frontal cortex in general, and it means a small or absent RewP in a patient with frontal damage should prompt a check of whether the lesion touches this right caudal region.

The priority for future work follows from the same asymmetry. The left caudal ACC question remains open because too few patients had lesions there, and the authors frame insufficient sample size as the major challenge of loss-of-function studies in this territory, where isolated anterior cerebral artery lesions are rare. Until a larger sample tests the left hemisphere, the defensible reading is a right-lateralized causal link in this cohort, with the single-generator claim resting on its convergence with the earlier intracranial evidence.

Sources used in this answer