THE FEEDBACK

By Joana Soldado-Magraner (CMU), Yuki Minai (CMU), Byron M. Yu (CMU), Matthew A. Smith (CMU)

Abstract

Delay period activity in the dorsolateral prefrontal cortex (dlPFC) has been linked to the maintenance and control of sensory information in working memory. The stability of working memory-related signals found in such delay period activity is believed to support robust memory-guided behavior during sensory perturbations, such as distractors. Here, we directly probed dlPFC’s delay period activity with a diverse set of activity perturbations and measured their consequences on neural activity and behavior. We applied patterned microstimulation to the dlPFC of two male rhesus macaques implanted with multielectrode arrays by electrically stimulating different electrodes in the array while they performed a memory-guided saccade task. We found that the microstimulation perturbations affected spatial working memory-related signals in individual dlPFC neurons. However, task performance remained largely unaffected. These apparently contradictory observations could be understood by examining different dimensions of the dlPFC population activity. In dimensions where working memory-related signals naturally evolved over time, microstimulation impacted neural activity. In contrast, in dimensions containing working memory-related signals that were stable over time, microstimulation minimally impacted neural activity. This dissociation could explain how working memory-related information may be stably maintained in dlPFC despite the activity changes induced by microstimulation. Thus, working memory processes are robust to a variety of activity perturbations in the dlPFC.

You can find this work by Joana and CMU colleagues at The Journal of Neuroscience here: Robustness of Working Memory to Prefrontal Cortex Microstimulation (2025).

Commentary by Joana Soldado-Magraner

We constantly integrate information from our environment to plan future behavior. To process and maintain this information in order to guide future actions, the brain must store it in working memory and protect it from interference from both external and internal signals. In this paper, we aimed to address the following question: to what extent is working memory robust to interference?

To answer this, we used electrical microstimulation to directly perturb working memory–related signals in the prefrontal cortex of non-human primates engaged in a memory-guided task. To our surprise, disruption of these signals did not impair the animals’ ability to perform the task. We thus concluded that working memory processes are robust to direct interference in the prefrontal cortex.

Our findings touch on one of the central philosophical tensions in neuroscience: the relationship between causal intervention and functional interpretation. At its core, our study challenges a seemingly intuitive assumption—that perturbing neural activity in a key cognitive region such as the prefrontal cortex should reliably disrupt the computations underlying working memory. Instead, the observed robustness invites a deeper reconsideration of how working memory is implemented and stabilized in the brain.

To resolve the apparent paradox, we adopted a computational perspective and interpreted the impact of perturbations from an information-encoding standpoint. We used statistical models to characterize how working memory information is represented within neural populations and how stimulation affects it. We found that this information is distributed across the population and remains accessible—or decodable—despite stimulation-induced changes in activity. This suggests that the prefrontal cortex may be endowed with mechanisms that encode and protect relevant information from external interference.

In addition to our proposed explanation, other robustness mechanisms may also be at play. Our findings are consistent with the idea that working memory is not primarily driven by localized patterns of activity within specific areas, but rather by distributed and redundant representations. From a philosophical perspective, this emphasizes that working memory “content” arises from patterns of interaction across networks spanning multiple brain areas. Accordingly, working memory representations may not be statically stored, but instead continuously regenerated through recurrent dynamics. Perturbations may transiently displace activity without fundamentally altering the underlying “attractor structure” that keeps activity “in place”—which may be sustained by both local and global recurrent interactions.

Alternatively, it is possible that our stimulation failed to impact behavior because it was delivered at a time when a motor plan had already been formed, or because the task was sufficiently easy that the induced changes were not strong enough to bias behavior. However, this still leaves open the question of how the neural system can operate reliably in the presence of such widespread changes in neural activity, as we observed.

More broadly, our study highlights the epistemological limits of causal perturbation as a tool for understanding brain function. In a system as complex as the brain, causality is not easily localized, and interventions can propagate in ways that are difficult to predict or measure. Although microstimulation has traditionally been considered a gold standard for establishing causality, its effects are notoriously complex. Stimulation does not simply “create” a given activity state in a vacuum; rather, it interacts with ongoing activity in nonlinear and state-dependent ways. The absence of a behavioral effect, therefore, does not straightforwardly imply the irrelevance of the targeted neurons. Instead, it may reflect redundancy in the system, distributed computations, or a misalignment between the perturbation and the relevant computational subspace—as we discuss in the paper.

Overall, our work not only advances the empirical understanding of working memory, but also prompts a broader reevaluation of how we conceptualize computation, representation, and causality in the brain.

Editor’s note: Please feel encouraged to share your thoughts and questions about this commentary and/or the article in the comment box below!