Left inferior frontal gyrus is critical for response inhibition

For more than a decade, the neuroscience community agreed on something: stopping yourself from doing something was a right-brain job — specifically, the right inferior frontal cortex. That was the story. Functional neuroimaging studies kept pointing there, the field kept repeating it, and it became a consensus that started to feel like a fact. Then Diane Swick, Victoria Ashley, and And U. Turken asked a question almost nobody had bothered to ask: what about the other side? Response inhibition is the ability to stop yourself from acting on an impulse — it’s the brake pedal of the motor system. It sounds simple, but it’s a fundamental executive function, and losing it has real consequences. The standard tools for studying it are the Go/NoGo paradigm and the Stop-Signal task. In the Go/NoGo version, you watch a rapid stream of letters and press a button for every letter except for the letter X. Most of the time, you press. The occasional NoGo trial is the moment that matters. And how often those NoGo trials appear turns out to be critical — because if stopping is rare, your brain builds a powerful habit of going, and overriding that habit becomes much harder. The consensus around the right inferior frontal gyrus — right IFG — was built on functional neuroimaging. Scan after scan, study after study, the right side lit up when people had to withhold a response. That evidence accumulated into something close to doctrine. What was conspicuously missing was systematic evidence about the left inferior frontal gyrus. This wasn’t because anyone had checked and found nothing — it was because almost nobody had checked at all. Swick and colleagues used a different approach entirely: neuropsychology. If a brain region is truly necessary for a function, then damage to it should impair that function. This is the logic of lesion studies. They recruited twelve patients with focal damage centered on the posterior left inferior frontal gyrus and insula — specifically Brodmann areas 44, 45, and 6, with some overlap into areas 9, 46, 47, and the temporal tip. The mean age of this group was fifty-eight point six years, with a mean of five point eight years since injury. Seven of the twelve had right-sided motor weakness, so they responded with their left hand — a detail that would matter when interpreting results. The comparison group was crucial: five patients with bilateral ventromedial orbitofrontal cortex damage. If you wanted to show that the left IFG deficit was specific, and not just a sign that any frontal damage disrupts behavior, you needed another frontal group — and the orbitofrontal patients provided exactly that control. Sixteen healthy adults, age-matched, rounded out the study. The task ran in four blocks of one hundred and forty trials each, alternating between two conditions. In the easy condition, half the trials were NoGo — stopping was required often, so it remained manageable. In the hard condition, only one in ten trials was NoGo. When stops are that rare, Go responses become nearly automatic. Your finger is already moving before the signal fully registers. That’s the condition where the inhibitory system really has to earn its keep. The results were unambiguous. Left IFG patients made significantly more false alarm errors — responses on trials where they should have stopped — than controls in both conditions. In the easy fifty percent block, the group difference was significant. In the hard ten percent block, the impairment was larger and more significant still. The interaction between group and NoGo probability was significant at a p-value less than point zero five, driven by that amplified deficit when inhibition was most demanding. The orbitofrontal patients, by contrast, performed like healthy controls in both conditions — not just similar, but statistically indistinguishable — and they were more accurate than the left IFG group, with that advantage especially clear in the hard condition. That contrast matters enormously. It rules out a narrative where any frontal damage degrades response inhibition. The deficit is specific to left IFG. There's a subtler finding in the reaction time data that clarifies what’s actually going wrong. On trials where patients made errors — pressing when they shouldn’t have — their responses were faster than on correct Go trials. That pattern held across all groups, with a highly significant main effect of accuracy. Swick and colleagues interpret this as impulsive responding: the signal to stop hadn’t been fully processed before the finger moved. These patients aren’t just slow to brake. They’re pressing before the braking even begins. Notably, there was no difference in error rates between patients who used their left hand versus their right hand, which means the deficit isn’t about hand dominance or motor coordination — it’s about the inhibitory process itself. Now here's where the study does something methodologically clever. Rather than simply challenge the imaging literature, the authors ran their own quantitative meta-analysis of it. Activation likelihood estimation — ALE — is a coordinate-based technique that pools activation foci from dozens of neuroimaging studies and identifies where they cluster consistently. The team searched BrainMap, pulled twenty-five studies directly from the database, and added fourteen more from PubMed, ultimately entering thirty-nine studies into the analysis. They used a full-width half-maximum of twelve, a false discovery rate threshold of a p-value less than point zero one, and a cluster extent cutoff of one hundred cubic millimeters. The ALE map produced thirteen significant clusters. The largest was centered in the right middle frontal gyrus and adjacent insular cortex, with additional peaks in the medial superior frontal gyrus — the pre-supplementary motor area — and right inferior parietal regions. Right-lateralized, yes. But that's not the whole picture. The analysis also identified a large cluster in the left insula extending into the putamen, and bilateral anterior insular involvement appeared consistently across the Go/NoGo and Stop-Signal literature. When you aggregate thirty-nine studies rather than cherry-picking the most dramatic ones, the imaging literature doesn’t tell a clean right-only story. It tells a bilateral story, with clear but not exclusive right-hemisphere dominance. That’s a very different map. This is the core of what Swick and colleagues argue — not that right IFG doesn’t matter, but that the field had been asking half the question. The lesion data shows left IFG is necessary. The meta-analysis shows it's active across the imaging literature when you look carefully enough. Both lines converge on the same place. The conclusion carries a methodological point worth pausing on. Neuroimaging and lesion methods are not interchangeable. Imaging shows you what regions are active; lesions show you what’s necessary. A region can be active without being critical — other areas might compensate. But when damage to a region consistently impairs a function, that’s harder to explain away. The left IFG result from this lesion study provides evidence that the imaging literature, read through a right-dominant lens, had been systematically underweighting. What remains open is exactly how left and right IFG divide the labor. They may perform identical computations in parallel, each capable of supporting inhibition independently. Or they may play distinct roles — different aspects of the stopping process or sensitivity to different task demands. The Stop-Signal and Go/NoGo paradigms may recruit the system differently, and the rarity of NoGo trials clearly modulates how hard the system has to work. Some of the earlier right-lateralization findings may reflect those task differences rather than a fundamental hemispheric asymmetry. But the practical implication Swick and colleagues leave you with is clear. The map of the inhibitory brain — the one that put right IFG at the center and left IFG largely off the page — needs revision. The left inferior frontal gyrus is not background circuitry. It's a critical node. And the way to see that was not to scan another hundred healthy people, but to ask what happens when it’s gone. This lecture was created by ennepō. Go to https://ennepo.ai to Discover, Create and Follow the latest research in your field. Read when you can. Listen when you want to.