Search Abstracts | Symposia | Slide Sessions | Poster Sessions
Left and right hemisphere language network functional connectivity across development
Poster Session C, Saturday, September 13, 11:00 am - 12:30 pm, Field House
Natalya Vladyko1, Peter Turkeltaub1,2, Elissa L. Newport1,2; 1Georgetown University Medical Center, 2MedStar National Rehabilitation Hospital
Language processing in healthy adults relies on a left-lateralized network. When the critical left hemisphere (LH) language network (LN) is damaged early in life, e.g., in perinatal stroke, language can develop in the right hemisphere (RH) without deficits (Newport et al.,2022). In contrast, stroke in adulthood is associated with limited recovery and persistent aphasia. What allows for such RH reorganization early in life but not later? How does communication between language-related areas within the LH and RH change over age? We hypothesized that strong LH language connectivity was not well-established in early life, and that its development over age should align with when RH reorganization after stroke can occur. We analyzed LN functional connectivity (FC) across four healthy age groups: infants (N=54, mean age=4 months), toddlers (N=61, mean age=38 months), children (N=111, mean age=10 years), and adults (N=71, mean age=61 years). Infant and toddler data was extracted from the Baby Connectome Project, children data from the Human Connectome Project Development. Adult data were collected at Georgetown University. Participants were scanned during rest or movie watching. After standard preprocessing, we computed FC between 246 regions defined by the Brainnetome atlas (Fan et al.,2016). We used the Neurosynth database to generate an activation map of studies that report a key phrase ‘language comprehension’. The LH parcels that overlapped with the map, and their RH homotopes were defined as the LN. All the parcels outside the LN were defined as a non-language network (NLN). To analyze the differences in FC between RH and LH we ran paired-sample t-tests at each edge value within the LN (91 total) and between the LN and NLN (2071 total). All results were corrected for multiple comparisons. Within the LN, adults showed a LH bias: of 91 LN edges, 26 were stronger in the LH and 9 in the RH (χ2(1)=10.9,p=.001). Similarly, children had 24 edges stronger in the LH and 9 in the RH (χ2(1)=6.82,p=.009). As hypothesized, this pattern of LH connectivity was not present in early life. Infants had no significant difference in the proportion of strong connections (4LH and 9RH, χ2(1)=1.9,p=.17). Similarly, toddlers showed no significant differences (6LH and 16RH, χ2(1)=4.6,p=.033; Bonferroni-corrected α-level across groups was .01). For the between-network FC, all age groups had a higher number of strong RH than LH connections (infants: 44LH, 112RH; toddlers: 49LH, 124RH; children: 238LH, 345RH; adults: 103LH, 174RH). This LN-to-NLN stability suggests that our LN changes are not due to methodological differences in FC data collection across samples and ages. Our results suggest a developmental shift in the number of strong LH LN connections. Early in life, LN connectivity is not yet adult-like: infants and toddlers have no difference in the strength of RH and LH LN connections. As the brain matures, FC within the LH LN increases, forming a more specialized and efficient network for language processing. We speculate that this changes the ability of the RH to develop language after stroke, a hypothesis we are pursuing in ongoing work on childhood stroke.
Topic Areas: Language Development/Acquisition,