Researchers Identify Speech-Intent Signals Outside the Frontal Lobe

Scientists are developing a brain-computer interface (BCI) that might help people with Broca's aphasia communicate. Broca's aphasia is a language disorder caused by damage to the frontal lobe: people may understand what they want to say but struggle to express it. New research found, for the first time, that the temporal and parietal cortices outside the frontal lobe also carry information about the intention to speak. That distinction matters for a future BCI intended to decode only speech a user wishes to express, without exposing thoughts they do not want to share.

When a word will not come

Imagine seeing a furry, four-legged animal that meows. You know what it is, yet cannot say the word “cat.” Broca's aphasia, also called expressive aphasia, affects speech or writing. Speech therapy is the main current treatment. Northwestern University researchers are exploring a possible alternative that would turn brain signals into language.

Before such an interface can be designed, researchers must determine where to record signals related to speech intention. Existing BCIs for people paralyzed by amyotrophic lateral sclerosis (ALS) or a brainstem stroke record frontal-lobe signals. Broca's aphasia, however, often follows stroke or a brain tumor that damages the frontal areas involved in language production. A BCI for these patients may therefore need signals from elsewhere in the brain.

The accompanying source image identifies the temporal and parietal cortices as regions involved in processing speech intention outside the frontal lobe. Image credit: Northwestern University.

Speech-intention signals outside the frontal lobe

Northwestern Medicine scientists found that the temporal and parietal cortices contain information related to the intention to speak. “It is a small but necessary step,” said corresponding author Marc Slutzky, a professor of neurology and neuroscience at Northwestern University's Feinberg School of Medicine. He said the finding could help distinguish a person's intention to speak from silence or private thought. The study was published on February 13, 2025, in the Journal of Neural Engineering.

Speech decoding and private thought

Slutzky said these early findings should help designers distinguish brain activity associated with producing language from activity associated with perceiving or understanding it. “From both a practical and an ethical standpoint, it is essential to avoid decoding thoughts the user does not wish to say,” he said. No treatment for aphasia was tested in this study.

Who took part

The researchers recorded activity in nine Northwestern Memorial Hospital patients without language impairment who were receiving care for epilepsy or brain tumors. They used implanted electrode arrays employed during presurgical epilepsy monitoring or electrodes temporarily placed in the operating room. Participants either read words shown on a screen aloud or remained silent while researchers recorded electrical signals from the cortical surface, known as electrocorticography (ECoG). The researchers plan to investigate whether the same approach can decode the words participants actually spoke.

Original report: SciTechDaily.