Back

Language areas revisited by principal component analysis of electrical cortical stimulation mapping

Otani, M.; Shimotake, A.; Nakae, T.; Sakamoto, M.; Matsuhashi, M.; Kikuchi, T.; Yoshida, K.; Kunieda, T.; Miyamoto, S.; Takahashi, R.; Lambon-Ralph, M. A.; Ikeda, A.; Matsumoto, R.

2025-03-21 neurology
10.1101/2025.03.21.25323362 medRxiv
Show abstract

Electrical cortical stimulation (ECS) remains the gold standard for language mapping in neurosurgery; thus, systematic and less invasive language tasks focusing on each language function should be developed. This study aimed to clarify the correlation between anatomy and language function by extracting principal components (PCs) directly related to language function from six tasks. Twelve patients with intractable temporal lobe epilepsy underwent ECS mapping with subdural electrode placement in the language-dominant hemisphere: 313 electrodes (3-48/patient) were confirmed to be located over the language areas by ECS mapping using six language tasks. Three major PCs were delineated; PC1, PC2, and PC3 represented reading, auditory receptive semantic processing, and expressive semantic processing, respectively. Anatomically, PC1 was prominent at the inferior frontal gyrus (IFG, especially pars opercularis), posterior parts of the superior temporal sulcus (STS), middle temporal gyrus (MTG), and inferior temporal gyrus; PC2 at the posterior part of IFG (extended more anteriorly than PC1), posterior superior temporal gyrus (STG), posterior MTG, and anterior ventral temporal cortex (VTC); and PC3 at the broad area including regions of the PC1 and PC2, and supramarginal gyrus (SMG) and VTC. Statistically significant functional differentiation was observed in the posterior language area (between the SMG and posterior STG) and basal temporal language areas (among the anterior, middle, and posterior parts). In conclusion, PC analysis revealed three independent language functions and indicated functional differences within the posterior and basal temporal language areas. Clinically efficient language mapping is expected by task selection, considering PCs and regions.

Matching journals

The top 7 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.