For decades, cognitive scientists have argued that what feels like “multitasking” is really the brain rapidly switching attention between tasks, one at a time, through a processing chokepoint in the prefrontal cortex. A new study from Georgetown University suggests that with enough practice, the brain can actually route a learned task around that chokepoint entirely, freeing up the prefrontal cortex to handle something else at the same time. The result, the researchers argue, is genuine parallel processing, not just faster switching.
The frontal bottleneck, and how to escape it
The study, titled “Extensive Experience Remodels Neural Task Circuitry to Escape the Frontal Bottleneck and Increase Automaticity of Categorization,” was published in the Journal of Cognitive Neuroscience in June 2026. It was led by Patrick Cox, PhD, now an assistant professor of psychology at Lehigh University, working with senior author Maximilian Riesenhuber, PhD, professor of neuroscience and co-director of the Center for Neuroengineering at Georgetown University School of Medicine, along with Georgetown co-authors Clara A. Scholl, Marissa L. Laws, Nelson E. Jaimes, and Xiong Jiang.
Using a dual-task paradigm built around car-category identification, the team put participants through more than 30,000 practice trials. Early on, brain imaging showed the classic bottleneck signature: activity concentrated in the prefrontal cortex, the region long understood to serve as a general-purpose hub for effortful, attention-demanding tasks, and a hard limit on how many such tasks a person could juggle at once. As training accumulated, that pattern shifted. The well-practiced categorization task migrated out of the prefrontal cortex and into the temporal cortex, the brain region associated with more automatic, pattern-based processing.
“The prefrontal cortex then stays free for whatever else you want to do, increasing your capacity,” Riesenhuber said, describing the effect of that handoff. Once a task is running on temporal-cortex circuitry instead of prefrontal circuitry, it stops competing for the same limited resource that a second, less-practiced task still needs, which is what allows the two to run in parallel rather than trading off.
Why this counts as real multitasking
The distinction the researchers draw is specific: task-switching means alternating full attention between two tasks quickly enough that it feels simultaneous, while what this study describes is two tasks genuinely sharing the brain’s processing capacity at the same time, because one of them no longer needs the resource the other is competing for. That reframes “multitasking ability” less as a fixed trait and more as something that extensive, task-specific practice can build, one well-learned skill at a time, by relocating it off the brain’s busiest circuitry.
An unusual pairing of funders
The research was supported by both the National Science Foundation (award BCS-1232530) and the U.S. Army Research Laboratory (award W911NF-24-1-0097), with additional support from the ARCS Foundation. The civilian-military combination reflects two very different institutional interests converging on the same basic-science question: NSF’s long-standing support for fundamental cognitive neuroscience, and the Army Research Laboratory’s applied interest in human performance under high cognitive load, relevant to training personnel who must reliably manage multiple demanding tasks at once.
What it means going forward
The findings add mechanistic detail to a question with obvious practical stakes, from cockpit and control-room design to how skills training is structured, by pointing to which tasks are good candidates for true parallel processing: not tasks a person is merely familiar with, but ones drilled to the point of automaticity. The authors’ framing suggests a testable principle for future work: multitasking capacity may be less about innate cognitive bandwidth and more about how much of what a person is doing has already been moved off the prefrontal cortex through practice.
Full details are available via Georgetown University School of Medicine’s press release on the study, published in the Journal of Cognitive Neuroscience.
