Strategic stabilization of arousal boosts sustained attention

Abstract

To meet their survival needs, organisms must continuously select which sensory stimuli to attend to and decide how much attention to pay. Attention’s selective aspect has been a cornerstone of behavioral and physiological study, whereas attentional intensity is poorly understood. Autonomic arousal is thought to strongly influence attentional intensity, but evidence is lacking, including regarding if and how organisms self-regulate their arousal to match attentional intensity to its utility. Here, we developed an auditory attentional intensity task for head-fixed mice and vary task utility by changing reward size. We record pupil size and walking speed as proxies of arousal and exploration, respectively. Using simple performance metrics and sequential sampling modeling, we find that mice increase their attentional intensity during periods of high task utility. The utility-related attentional boost is partially mediated by stabilization of pupil-linked arousal around an optimal mid-sized level, and an associated reduction in exploratory behavior. In sum, self-regulation of arousal partially implements strategic attentional intensity allocation.

A disinhibitory circuit mechanism explains a general principle of peak performance during mid-level arousal

“…Next, we performed three additional analyses to test whether the quadratic arousal-performance relationship was driven by significant behavioral or pupil events on the previous trial. Specifically, 1) we excluded all trials that followed erroneous responses (23.1% of all trials), possibly upregulating arousal, 2) we performed the analyses after regressing out task-evoked pupil responses on the previous trial (de Gee et al, 2014Gee et al, , 2022Waschke et al, 2021), and 3) we tested whether the quadratic arousal-performance relationship would also be revealed if the task-evoked pupil response in the previous trial was used as the measure of the pupil-linked arousal state. These analyses further solidified our main conclusions and indicated that the inverted Ushaped relationship between pre-stimulus (baseline) pupil size was not driven by response errors or task-evoked pupil responses on the previous trial (Supplementary Figure 3).…”

“…In fact, although the observed quadratic arousalsensitivity relationship was clearly robust to task parameters, some previous studies in humans have also reported linear effects. In mice, primarily U-shaped functions are observed (de Gee et al, 2022;McGinley, David, et al, 2015;McGinley, Vinck, et al, 2015): mice perform best at an auditory detection task during quiet wakefulness, and worse during states of low and high arousal (including locomotion) (McGinley, David, et al, 2015). In humans however, the shape of the arousal-performance relationship appears to be less clear.…”

Sensory processing in humans and mice fluctuates between external and internal modes

“…While the increase in response speed and history congruence over time (S1 Text) may argue against insufficient training as an alternative explanation for internal mode processing, it may also be indicative of waning arousal. The multiple mechanistic mappings to RTs and confidence warrant more direct measures of PLOS BIOLOGY arousal (such as pupil size [28,[63][64][65]67,68], motor behavior [63,68], or neural data [69]) to better delineate bimodal inference from fluctuations in global modulators of task performance.…”

Foraging Under Uncertainty Follows the Marginal Value Theorem with Bayesian Updating of Environment Representations

“…Tone cloud stimuli consisted of repeating chords divided into 20 ms bins [de Gee et al, 2022]. Each chord was comprised of 5 semitones randomly selected between 1.5 kHz and 96 kHz.…”