Predicting hemispheric dominance for language production in healthy individuals using support vector machine.

Area of Science:

• Neuroscience
• Cognitive Neuroscience
• Neuroimaging

Background:

• Hemispheric specialization for language is a fundamental concept in neuroscience.
• Understanding language dominance patterns in non-typical individuals is crucial for clinical and research applications.
• Functional laterality index (HFLI) using fMRI-BOLD is a key metric for assessing language lateralization.

Purpose of the Study:

• To classify hemispheric language dominance patterns in individuals with non-typical language organization using a Support Vector Machine (SVM) classifier.
• To investigate the prevalence of different brain patterns (left-dominant, right-dominant, co-dominant, co-non-dominant) in this population.
• To compare language dominance patterns between non-typical right-handers (RH) and left-handers (LH).

Main Methods:

• A Support Vector Machine (SVM) classifier was trained on fMRI-BOLD data from 250 typical individuals to discriminate language dominance.
• The trained SVM was applied to assess hemispheric language activation patterns in 47 non-typical individuals based on their HFLI.
• Analysis focused on sentence minus word-list production fMRI-BOLD contrast maps.

Main Results:

• The SVM classifier identified at least one dominant hemisphere in all but 3 of the 47 non-typical individuals, suggesting dominant organization is most frequent.
• Left hemisphere dominance was predicted in all non-typical RH and 57.4% of non-typical LH.
• The most frequent pattern was left-dominant (51%), followed by right-dominant (23%) and co-dominant (19%). RH predominantly showed left-dominant or co-dominant patterns, while LH exhibited all types.

Conclusions:

• Dominant hemispheric language organization is the most common pattern, even in individuals with non-typical language profiles.
• Left-handers display a wider spectrum of language organization patterns compared to right-handers.
• The findings highlight the adaptability of brain functional organization for language across different individual profiles.