Skip to Content

One brain region, two tasks: how ACC outputs divide social behavior and attention

Science Advances · 2026-07-08

Table of contents

No titles in the blog post

Science Advances · Research frontiers

One brain region, two tasks: how ACC outputs divide social behavior and attention

A mouse study finds that anterior cingulate outputs to zona incerta and limbic thalamic reticular nucleus separately modulate partner interaction and attention-task performance.

AI scientific concept illustration inspired by the research theme

Interacting with a partner recruits attention, while an attention task involves waiting, choosing and controlling responses. How does the brain divide work across these related behaviors? A 2026 Science Advances study examines the anterior cingulate cortex, or ACC, a region often implicated in both social behavior and attention. Projection mapping, cellular characterization and pathway modulation reveal two minimally overlapping populations targeting zona incerta, or ZI, and limbic thalamic reticular nucleus, or lTRN. Multiple functions within one region thus acquire a more detailed cellular organization.

KEY FINDING

In mouse tasks, ACC→ZI mainly modulates partner investigation, while ACC→lTRN modulates attention-task performance and premature responding. Enhancing each output improves the corresponding behavior in Shank3b-deficient mice.

Contents · 6 sections

Social interaction and correct choices both recruit ACC

CaMKII-positive ACC population calcium increased during partner contact and decreased during retreat, without matching changes during digging, rearing or grooming. In the five-choice serial reaction task, mice waited for a visual cue and selected the correct aperture. ACC signals rose during waiting and cue phases, mainly on correctly completed trials. The region was recruited in both kinds of task.

Suppressing the broad ACC excitatory population reduced partner investigation and correct performance and increased premature responding, while measures including locomotion and object memory remained similar. Regional effects were observed in both sexes, with sex differences in some attention and premature-response measures. Shared recruitment then led to a deeper question: which outputs perform the work in each task?

Two populations connect with different targets

Activity-dependent labeling outlined two task-associated output patterns: socially recruited ACC cells prominently targeted ZI, whereas attention-recruited cells prominently targeted lTRN. Dual retrograde labeling showed about 1.61% overlap between the populations. Output targets thus identified largely separate populations with a small shared component.

The map made ACC function more concrete. Neurons within the same cortical region may participate in different behaviors through their distinct connections. Comparing cellular properties and then observing behavior during modulation of each output connected the anatomical organization with functional specialization.

Original figure 3Task-associated outputs and population overlap
Task-associated outputs and population overlap
Original Figure 3. A–F show social- and attention-associated outputs; G–K show overlap between ACC projection populations using dual-target labeling, outlining their anatomical distribution.
Source: Wang F et al. · Science Advances · CC BY-NC
Original figure and English caption ↗

Different outputs have different cellular profiles

The populations had similar resting potentials, but lTRN-projecting cells had lower rheobase, higher membrane resistance and greater firing under stimulation. ZI-projecting cells had longer, more complex dendrites. Electrical and structural features thus showed cellular differences accompanying the projection organization.

Optically evoked excitatory currents in target regions were sensitive to synaptic blockade, supporting functional glutamatergic connections. Transcript analysis also identified differentially expressed genes, adding molecular features to the two populations. Their dendrites, firing and expression profiles offer leads for studying how each participates in network computation.

Original figure 4Multilevel features of projection cells
Multilevel features of projection cells
Original Figure 4. B–E show electrophysiology; F–I show dendritic morphology; J–M show transcriptional analysis and pathway enrichment, comparing the projection populations.
Source: Wang F et al. · Science Advances · CC BY-NC
Original figure and English caption ↗

The ZI output influences partner investigation

Suppressing ACC terminals in ZI reduced sniffing of a mouse-containing cup, sociability indices and contact during free interaction. Suppressing ACC→lTRN did not significantly change the corresponding social readouts. The two outputs affected the same behavioral questions differently, placing social specialization within a defined projection.

Light controls and baseline behavior measures helped relate the changes to projection activity rather than a general reduction in movement. Convergent effects across partner-investigation tasks clarified the ZI output’s contribution. The investigation then turned to attention tasks to ask whether the other output performed different work.

Original figure 5Two outputs and partner investigation
Two outputs and partner investigation
Original Figure 5. E–H show reduced partner investigation after ACC→ZI inhibition; I–L show nonsignificant corresponding changes after ACC→lTRN inhibition. These experiments used male mice.
Source: Wang F et al. · Science Advances · CC BY-NC
Original figure and English caption ↗

The lTRN output influences task performance

In the five-choice task, ACC→lTRN inhibition reduced correct performance and increased premature responding. Related changes appeared with modulation across the whole trial or restricted to waiting or cue phases. ACC→ZI inhibition did not significantly change these measures. Together with the social findings, the outputs showed a double dissociation across the tasks.

The task combines waiting, visual selection and response control. The lTRN output therefore relates to how animals use cues and regulate response timing. Behavioral specialization aligned with anatomical organization, providing direct clues to how the ACC organizes behavior through different connections.

Original figure 6Two outputs and the attention task
Two outputs and the attention task
Original Figure 6. A–L show ACC→lTRN inhibition at different phases; M–X show corresponding ACC→ZI comparisons, assessing correct performance and premature responding in males.
Source: Wang F et al. · Science Advances · CC BY-NC
Original figure and English caption ↗

Enhancing each output improves its corresponding behavior

Shank3b-deficient mice showed both social and attention-task deficits, with attention problems more evident later in the task. Both ACC populations had altered excitatory inputs and dendritic spines, with differing details across ZI and lTRN connections. Enhancing ACC→ZI improved social behavior, while enhancing ACC→lTRN improved correct performance, extending the parallel-output findings to a genetic model.

Most projection-behavior and model-improvement experiments used males, while regional interventions and some cellular analyses included both sexes. Full specialization across sexes and coordination of social behavior and attention during natural interaction will extend the work. The identified cellular and connection organization offers a more detailed, testable account of multiple functions within one region.

Original figure 8Projection modulation in the genetic model
Projection modulation in the genetic model
Original Figure 8. E–H show social behavior after ACC→ZI activation; I–T show attention-task performance after ACC→lTRN activation in male Shank3b-deficient mice.
Source: Wang F et al. · Science Advances · CC BY-NC
Original figure and English caption ↗

Publication details

Wang F, Chen M, Lyu M, Zhang Y, Sun H, Lu Y, Cui D, Xu W, Lyu Y, Zhai Y, Han J, Li J, Xu H, Wang W, Chen L, Ma X, Li Y.
Parallel anterior cingulate cortex pathways underlie dissociable social and attentional outputs
Science Advances · 2026-07-08
DOI: 10.1126/sciadv.aec8032