Back

Bayesian analysis of resting-state temporal low-γ power in children with phonemic decoding deficits and certain comorbidities

Lasnick, O. H. M.; Hancock, R.; Hoeft, F.

2023-09-22 neuroscience
10.1101/2023.09.20.558564 bioRxiv
Show abstract

Dyslexia (decoding-based reading disorder, or RD) is a common learning disorder affecting a large proportion of the population. One theory of the origins of reading deficits is a language network which cannot effectively entrain to speech, with cascading effects on the development of phonological skills. Low-gamma (low-{gamma}, 30-45 Hz) neural activity is thought to correspond to tracking at phonemic rates in speech. The main goals of the current study were to investigate temporal low-{gamma} band-power during rest in a sample of children and adolescents with and without RD. We used a Bayesian statistical approach, which has become increasingly popular in the field for its ability to quantify the relative likelihood of competing hypotheses. We examined whether (1) resting-state temporal low-{gamma} power was attenuated in the left temporal region in those with RD; (2) low-{gamma} power covaried with individual performance in reading skills; (3) low-{gamma} temporal lateralization was atypical in the group with RD. Results did not support the hypothesized effects of RD status and poor phonemic decoding ability on left hemisphere low-{gamma} power or lateralization: post-hoc tests revealed that the lack of atypicality in the RD group was not due to the inclusion of those with comorbid attentional deficits. However, post-hoc tests also revealed a specific left-dominance for low-{gamma} rhythms in the RDs only when participants with attentional deficits were excluded. We also observed an inverse relationship between phonemic decoding and left-lateralization in the controls, such that those with better decoding skills were less likely to show left-lateralization, contrary to our expectations. We discuss these unexpected findings in the context of prior theoretical frameworks on temporal sampling, and suggest that our null findings in resting-state EEG may reflect the importance of real-time language processing to evoke gamma rhythms in the phonemic range during childhood and adolescence.

Matching journals

The top 10 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.