Slower prefrontal metastable dynamics during deliberation predicts error trials in a distance discrimination task

Abstract

Previous studies have established the involvement of prefrontal cortex (PFC) neurons in decision processes in many task contexts. Single neurons and populations of neurons have been found to represent stimuli, actions, and internal deliberations. However, it is much less clear which underlying computations are affected during errors. Neural activity during errors can help to disambiguate confounds and clarify which computations are essential during a specific task. Here, we used a hidden Markov model (HMM) to perform a trial-by-trial analysis of ensembles of simultaneously recorded neurons from the dorsolateral prefrontal (PFdl) cortex of two rhesus monkeys performing a distance discrimination task. The HMM segments the neural activity into sequences of metastable states, allowing to link neural ensemble dynamics with task and behavioral features in the absence of external triggers. We report a precise relationship between the modulation of the metastable dynamics and task features. Specifically, we found that errors were made more often when the metastable dynamics slowed down, while trial difficulty influenced the latency of state transitions at a pivotal point during the trial. Both these phenomena occurred during the decision interval and not following the action, with errors occurring in both easy and difficult trials. Thus, modulations of metastable dynamics reflected a state of internal deliberation rather than actions taken or, in the case of error trials, objective trial difficulty. Our results show that temporal modulations of PFdl activity are key determinants of internal deliberations, providing further support for the emerging role of metastable cortical dynamics in mediating complex cognitive functions and behavior.

Metastable dynamics of neural circuits and networks

“…4.1 Hidden states coding for sensory, motor and cognitive variables Since these original works, metastable dynamics has been reported in rat gustatory cortex (GC) (Jones et al, 2007), in monkey somatosensory, motor, and premotor cortex (Kemere et al, 2008;Ponce-Alvarez et al, 2012;Mazurek et al, 2018), in monkey's area V4 (Engel et al, 2016), in monkey's orbitofrontal (Rich and Wallis, 2016), parietal (Bollimunta et al, 2012) and dorsolateral prefrontal cortex (Benozzo et al, 2021), in the hippocampus of rats (Maboudi et al, 2018), in the zebrafish (Marques et al, 2020) and in multiple human brain areas (Taghia et al, 2018).…”

“…In addition to sensory information, metastable activity has been implicated in changes of behavioral and cognitive states during tasks requiring attention (Engel et al, 2016), expectation (Mazzucato et al, 2019), decisions (Ponce-Alvarez et al, 2012;Rich and Wallis, 2016;Sadacca et al, 2016;Benozzo et al, 2021), spatial navigation (Maboudi et al, 2018;Marques et al, 2020) and working memory (Ponce-Alvarez et al, 2012;Taghia et al, 2018).…”

“…Coding states in the rat GC have been found to code for specific taste stimuli (Mazzucato et al, 2019). More recently, states coding for more abstract stimulus features have been found in a distance-discrimination task in which monkeys had to report which of two stimuli were farther from a central location on a computer screen (Benozzo et al, 2021). In this case, it was found that some states reflect the relative distance of the two stimuli based on their features (e.g., whether they were a blue circle or a red square, see Fig.…”

“…These have been dubbed "coding states", as they convey information about the stimulus (Mazzucato et al, 2019). Metastable states have been found to code for more abstract concepts such as the relative distance of two target stimuli based on stimulus features (Benozzo et al, 2021). Besides the meaning of coding states, it is their organization in sequences that promises the largest benefit in terms of coding.…”

“…For instance, in GC, coding states for specific tastants tend to be within the first 0.5 s following stimulus presentations, but shift towards earlier onset times in trials when a stimulus is expected -providing a potential neural substrate of expectation (Mazzucato et al, 2019). On the other hand, monkeys performing a distance-discrimination task tend to make errors when state sequences stretch out in time, i.e., when the metastable dynamics slows down (Benozzo et al, 2021).…”