Dynamic states of population activity in prefrontal cortical networks of freely-moving macaque

1) Research question and significance

• The study investigates how population activity in a high-level executive circuit tracks behavioral state in freely moving primates, and whether cortical state dynamics extend beyond sensory regions to areas such as the dorsolateral prefrontal cortex (dlPFC).

• By linking arousal, attention, and cognitive control to real-time population patterns, the work provides insight into how cortical state supports goal-directed behavior in a naturalistic setting.

2) Methods and data (including the wireless technology, eye tracking, video analysis, subject model, and brain region)

• An integrated wireless system recorded brain activity, eye movements, and behavior in two macaques. A 96-channel Utah array was implanted in the left dlPFC (area 46) to capture spikes and local field potentials (LFPs); a battery-powered headstage conveyed signals to a central processor via eight directional antennae.

• A wireless eye tracker monitored pupil diameter and oculomotor activity; overhead video classified behavior into active wakefulness, quiet wakefulness, or rest. Rest required a sustained posture; wake epochs were subdivided using a motion metric derived from frame-to-frame image differences.

• Neural analyses yielded a population synchrony index (PSI), derived from the 0.5–10 Hz content of the population firing rate in 10 s epochs. Units were categorized as narrow-spiking or broad-spiking based on spike waveform to approximate interneurons versus pyramidal cells.

3) Key findings (cortical substates, transitions, neuron-type differences)

• Rest produces a synchronized cortical state: higher low-frequency power, larger PSI, and reduced mean population firing rate.

• Wakefulness desynchronizes the cortex; active wakefulness shows higher firing rates and lower PSI than quiet wakefulness, with quiet wakefulness intermediate between rest and active wakefulness.

• PSI outperforms the LFP power ratio in predicting behavioral state (linear SVM decoding).

• Narrow-spiking neurons exhibit stronger population synchrony than broad-spiking neurons across states; rest shifts the network toward inhibition (lower E/I ratio). During wakefulness, arousal driven by locomotion and pupil dilation reduces slow fluctuations, especially in the narrow-spiking subpopulation.

• Behavioral metrics correlate with cortical state: motion index and pupil diameter relate to firing rate and PSI. Specifically, motion correlates positively with firing rate (R ≈ 0.44) and negatively with PSI (R ≈ -0.12); pupil diameter correlates positively with firing rate (R ≈ 0.45) and negatively with PSI (R ≈ -0.27).

• Rest, wake, and transitions show persistent PSI changes over minutes, suggesting slow, arousal-linked modulation of population dynamics. The 2D electrode geometry reveals higher PSI at nearby cortical distances, though state differences persist across distances.

4) Mechanistic interpretation and implications (how behavior and arousal modulate cortical state)

• Cortical state reflects a dynamic interplay between behavioral demands and arousal rather than a single desynchronized wakeful mode. Arousal signals (e.g., acetylcholine/norepinephrine release) associated with movement and pupil dilation modulate ongoing slow fluctuations, biasing networks toward desynchronization during active engagement.

• Putative parvalbumin interneurons (narrow-spiking) appear central to regulating low-frequency synchrony, potentially constraining excitatory drive and stabilizing activity during behavior.

5) Limitations and potential future directions

• Eye-tracking resolution excludes microsaccades; recordings are confined to a single cortical area (dlPFC), limiting cross-regional inferences.

• Cell-type classification relies on waveform shape, not genetics; laminar specificity is not resolved with the current chronic array.

• Future work could expand to multiple frontal areas, laminar probes, cross-area communication analyses, and causal manipulations of arousal systems.

6) Overall synthesis and potential applications or impact

• The work demonstrates a scalable, wireless approach to study cortical states in natural behavior and shows that executive circuits exhibit state-dependent population dynamics tightly linked to arousal and motor activity.

• Findings implicate inhibitory (narrow-spiking) circuits in shaping population synchrony and suggest novel targets for understanding attentional control and cognitive disorders where arousal-state regulation is disrupted.