Decoding how the brain understands sentences in real-time

In a recently published study, Dr Nature communication, Researchers examined the neural networks in the left language-dominant hemisphere responsible for semantic integration using intracranial recordings in epilepsy patients during a reading task. These tests were also used to distinguish between effects driven by semantic coherence and task-based referentiality.

Study: Spatiotemporal dynamics of semantic integration in the human brain. Image credit: Triff/Shutterstock.com

Background

Understanding the neural mechanisms responsible for human sentence processing is critical to understanding the structure and timing of cortical computations. Despite the fundamental role of language in cognition, enabling us to extract meaning from unfamiliar cues, the specific brain regions responsible remain controversial.

Certain areas, such as the posterior temporal lobe, as well as the prefrontal and parietal cortices, are involved in language development; However, the consensus among researchers remains controversial. Furthermore, traditional methods such as functional magnetic resonance imaging (fMRI) lack the resolution to dissect the complexities of language processing.

Thus, additional research is essential to better understand specific cortical regions and their specializations in semantic processing, especially given the variability in the literature and the challenges of isolating distinct semantic processes with high spatiotemporal resolution.

About the study

In the present study, 58 native English-speaking patients aged 18 to 41 years were tested after providing informed consent. Patients with significant neurological history or anomalies such as prosopagnosia were excluded.

All study participants underwent thorough neuropsychological evaluation. All procedures were approved by the University of Texas Health Science Center at Houston’s Committee for the Protection of Human Subjects.

Data were collected using depth or subdural grid electrodes that were surgically implanted, and their positions were confirmed using a combination of MRI and computed tomography (CT) imaging. Intracranial data collection began one day after electrode implantation for depth electrodes and two days for grid electrodes. Data were checked for noise and artifacts and any unreliable data points were discarded.

For the main test, patients were shown words and asked to immediately and accurately name everyday objects based on those words. Some sentences were designed to be incoherent, thus challenging patients to determine meaning. A secondary normative study was conducted on a non-clinical population to confirm the effectiveness of the stimulation.

A thorough analysis of the collected data revealed that of over 13,000 electrode contacts implanted, only 9,388 were considered suitable for analysis. The raw data were subsequently filtered, transformed and smoothed to discern important activation points during the trial.

The data were then mapped onto a cortical surface model to understand its significance. A significant portion of the analysis was devoted to understanding the interrelationship between language systems and episodic memory networks. The study also used the Human Connectome Project to determine regions of interest and ensure precise electrode placement.

Study results

The average individual reaction time after the final word in a sentence was 1,765 milliseconds (ms). When individuals encountered referential trials, their utterance reaction times were significantly faster than on non-referential trials.

Advanced mapping techniques were used to gain insight into the spatiotemporal dynamics of orthographic sentence processing. When reading a sentence, there was a gradual increase in activation in certain brain regions, particularly the inferior frontal gyrus, medial parietal cortex, anterior temporal lobe, and posterior middle temporal gyrus.

Later stages of sentence processing revealed activation in the ventromedial prefrontal cortex, posterior cingulate, and orbitofrontal cortex. Some regions remained active and showed higher activity towards the end of the sentence reading process.

At the onset of the final word, reference trials demonstrated greater brain activity in several areas, such as the middle frontal gyrus (MFG), middle inferior frontal sulcus (IFS), medial parietal cortex (MPC), parahippocampal cortex, ventromedial prefrontal cortex (vmPFC), and orbitofrontal cortex. (OFC). In comparison, non-referential trials elicited increased activity in the posterior superior temporal cortex and anterior inferior frontal gyrus immediately after the onset of the final word.

When exploring semantic coherence, distinct brain activity patterns were observed in the analysis of non-referential sentences based on their coherence. Incongruent non-referential sentences cause heightened activity in the medial frontal cortex and superior medial parietal cortex. Congruent non-referential sentences resulted in increased activity in regions such as the IFS, anterior inferior frontal gyrus (aIFG), angular gyrus, posterior middle temporal gyrus (pMTG) and OFC.

Further studies examine integrative lexical access by assessing semantic narrowness, which refers to the likelihood of identifying a defined object even before the final word of a sentence is presented. For example, while some sentences gave clear hints about the object in question, others were ultimately ambiguous.

There was no significant difference in articulation reaction times between the strong and limited constriction conditions. Additionally, there were no significant differences in sentence length or final word frequency between the two conditions.

Referential trials with limited semantic narrowing increased brain activity in regions such as the posterior superior temporal sulcus, MPC, IFS, anterior temporal lobe, and OFC, particularly after final word onset.