You are halfway through a job interview, reaching for the right word, and it slips out: “um.” The hiring manager’s eyebrow twitches. In that fraction of a second, you have been silently downgraded from “confident candidate” to “unprepared rambler.” But a growing body of peer-reviewed neuroscience suggests the person judging you has it backward. That tiny syllable is not a sign of mental vacancy. It is a byproduct of your brain doing some of its most sophisticated work in real time.
Multiple independent research groups have now linked filled pauses, the “uhs” and “ums” scattered through everyday speech, to executive brain function: the neural machinery responsible for planning, monitoring, and organizing language before it leaves your mouth. As of June 2026, the evidence spans behavioral experiments, functional MRI, and direct electrical recordings from the brain’s surface, and it consistently points in the same direction.
The behavioral evidence: fillers track with cognitive ability
One of the clearest demonstrations comes from a 2013 study published in Acta Psychologica. Researchers had 106 participants complete standardized cognitive assessments alongside a controlled sentence-production task. They cataloged multiple types of disfluency, including filled pauses like “uh” and “um,” and found that these speech patterns correlated with measurable differences in both executive control and general intelligence.
The key finding was not that smarter people say “um” more or less often. It was that the pattern and placement of fillers varied systematically with how well each person managed the cognitive demands of assembling sentences on the fly. The disfluencies were not random noise. They tracked with the brain’s ability to juggle competing demands: retrieving vocabulary, structuring grammar, monitoring for errors, and keeping the overall message on course, all simultaneously.
What brain imaging reveals
Neuroimaging studies have added physiological weight to those behavioral findings. In one fMRI experiment, researchers scanned patients with formal thought disorder, a condition seen in schizophrenia that involves well-documented impairments in executive control and language monitoring, while those patients described Rorschach inkblots aloud. The study connected pause behavior and filled pauses to discourse planning and monitoring processes in the brain. When executive systems were compromised, the pattern and frequency of pauses shifted in measurable, predictable ways.
Separate work on healthy speakers tells a complementary story about the brain regions that keep speech on track. A 2019 study published in Nature Communications reported that activation in the inferior prefrontal cortex, specifically Brodmann area 45, predicted whether extended speech stayed coherent or drifted off topic. Importantly, this study focused on discourse coherence rather than filler production specifically, but it identified the same executive-control regions in prefrontal cortex that other research has linked to speech planning. The researchers concluded that these regions support goal-directed speech production at the discourse level. When prefrontal activity dipped, speakers lost the thread of what they were saying. The overlap between the neural systems that maintain coherence and those implicated in filler production suggests a shared executive-control foundation, though the two phenomena are not identical.
The most granular evidence comes from electrocorticography, a technique that places electrodes directly on the brain’s surface. In epilepsy patients who had electrodes implanted for clinical monitoring, researchers recorded neural activity during naturalistic sentence production and compared what happened in the milliseconds surrounding a filler word versus an ordinary word. They found greater high-gamma activity, neural oscillations in the 70 to 110 Hz range associated with active cortical processing, for fillers in association cortex and visual cortex. Producing an “uh” or “um” recruited brain areas involved in higher-order processing, not just the motor regions that move the tongue and lips. (This finding has been reported in electrocorticography research on speech production, though no publicly available URL for the specific study can be confirmed at this time.)
Fillers help listeners, too
The cognitive work behind a filler does not only benefit the speaker. Research published in PLOS ONE demonstrated that hearing “um” before a word actually facilitated recognition of the word that followed. Both “um” and matched silent delays produced this effect, consistent with a delay mechanism rather than a pure error signal. The pause gives the listener’s brain extra processing time, and the filler itself may function as a cue that something less predictable is coming next, prompting the listener to pay closer attention.
This aligns with an influential analysis published in Cognition by psycholinguists Herbert Clark and Jean Fox Tree, who argued that “uh” and “um” are conventional English-language signals with systematic usage in spontaneous speech, not meaningless vocal garbage. Their work reframed fillers as part of the communicative toolkit rather than failures of it.
Where the science gets murkier
The connection between filled pauses and executive function is well supported, but several important questions remain open.
Aging. No published neuroimaging studies have directly examined how filler use changes in neurotypical aging populations. Researchers have studied clinical groups like patients with formal thought disorder, but whether the same cortical patterns hold for healthy older adults experiencing age-related cognitive shifts has not been experimentally tested.
Clinical applications. Speech therapists and communication coaches have long treated fillers as habits to eliminate. But no published intervention studies have examined how these neuroscience findings should reshape clinical practice. The gap between laboratory evidence and real-world speech therapy protocols is significant. It is possible that some fillers reflect helpful planning processes while others signal monitoring breakdowns, but current data do not support those fine-grained clinical distinctions yet.
Language and culture. The studies cited here were conducted primarily with English-speaking participants. Whether “uh” and “um” carry the same cognitive and communicative weight in Mandarin, Arabic, Swahili, or structurally different languages is an open question. Clark and Fox Tree’s analysis, for instance, was grounded entirely in English-language data.
Individual differences. Personality traits, anxiety levels, and speaking experience could all influence how often someone uses fillers and where those fillers land in a sentence. The existing experiments were not designed to disentangle these variables, leaving open the possibility that multiple psychological factors converge on similar disfluency patterns. Gender differences in filler use, documented in some sociolinguistic research, also remain largely unexamined from a neuroscience perspective.
Correlation, not causation, and why that still matters
Readers should be careful not to overinterpret these findings. The behavioral study found that speech disfluencies tracked with cognitive measures, but it did not establish that executive function directly causes specific pause patterns. The neuroimaging work shows that certain brain regions are more active during filler production, which is consistent with executive involvement but does not rule out contributions from lexical difficulty, emotional arousal, or social pressure. The electrocorticography researchers themselves emphasized that their findings identify neural signatures, not a complete mechanistic model.
Context matters, too. The fMRI study of patients with formal thought disorder examined a population with established executive-control impairments, making it a useful test case but limiting how directly the results generalize to everyday speakers. The healthy-speaker fMRI work focused on specific discourse tasks in controlled laboratory settings that differ from the messy reality of spontaneous conversation, with its interruptions, overlapping talk, and shifting goals.
There is also the problem of lumping all fillers together. An “um” at the start of a presentation, an “uh” in the middle of a complex explanation, and a trailing filler at the end of a sentence may serve different functions. Some may reflect strategic planning, others a search for the right word, and still others a reaction to social discomfort. Most studies aggregate these events into a single measure, which is useful for detecting broad patterns but risks obscuring subtypes that listeners intuitively recognize.
What this means the next time you catch yourself saying “um”
For anyone who has ever been told to eliminate fillers from their speech, the practical takeaway is straightforward: occasional “uhs” and “ums” are not reliable signs of incompetence or dishonesty. Within the limits of current evidence, they appear to be part of how the brain manages the demanding, real-time task of converting thought into fluent language.
That does not mean more fillers are better, or that sprinkling in extra “ums” will sharpen your thinking. The research suggests fillers arise naturally from the interaction between executive control, linguistic planning, and conversational context. They are neither mere bad habits nor secret markers of brilliance. They are something more interesting: small, audible windows into the brain’s continuous effort to keep language, thought, and social interaction running at the same time. The next time someone pauses mid-sentence and mutters “um,” what you are hearing is not a malfunction. It is the sound of a brain working hard.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
