Children diagnosed with self-limited epilepsy with centrotemporal spikes exhibit elevated functional connectivity between motor cortices and language regions during active language processing, and this heightened neural synchrony correlates with lower phonological performance. In an observational neuroimaging investigation conducted by researchers at Stanford University, high-density electroencephalography demonstrated that pediatric patients engaged in silent word processing generated abnormally high phase synchronization across distant brain networks. Rather than showing localized disruptions or diminished communication, children with this common childhood focal epilepsy demonstrated widespread hyperconnectivity across cortical areas responsible for motor output and speech comprehension.
However, clinical testing consistently documents mild to moderate cognitive deficits in affected children, who frequently lag up to a full standard deviation behind their typically developing peers on standardized assessments of expressive language and phonological awareness. While nocturnal epileptiform spikes in motor areas have long been suspected of contributing to these functional deficits, spike frequency itself does not reliably predict the magnitude of a child’s linguistic delay. Consequently, investigators hypothesized that the functional connections linking motor cortices to the developing language network undergo chronic, pathological alterations that impair cognitive efficiency.
To investigate how functional connections operate during active cognitive demands, the research team recruited thirty-one children with clinical diagnoses of the epilepsy syndrome and thirty-two typically developing control participants matched for age and sex. All participants were between five and thirteen years old, right-handed, and born after thirty-five weeks of gestation, with children presenting major neurological traumas or severe systemic conditions excluded. Demographic evaluations confirmed equivalent age distributions, with the patient group averaging 9.7 years and controls averaging 9.1 years, alongside comparable sex distributions and handedness scores. Cognitive evaluations using the Wechsler Abbreviated Scale of Intelligence and the Comprehensive Test of Phonological Processing revealed significant differences between groups: pediatric patients exhibited lower mean scores in full-scale intellectual quotients and phonological awareness composites compared to controls, alongside slightly lower socioeconomic index scores.
The experimental protocol recorded high-density electroencephalography across 128 scalp channels as children sat in a sound-attenuated environment facing a computer display and speakers. Participants completed three distinct randomized cognitive blocks: an expressive verb generation task, an overt word repetition task, and an eyes-open resting task. The verb generation task presented auditory recordings of concrete nouns chosen from standardized word databases, prompting participants to mentally generate an associated action word during a three-second covert processing window before an overt vocalization cue. The repetition task employed matching nouns but required participants only to retain the word for subsequent overt recitation, controlling for working memory and motor planning without engaging semantic retrieval networks. The passive resting baseline required participants to visually maintain fixation on a central screen cross without language cues.
Neural signals were preprocessed using the Maryland Analysis of Developmental EEG pipeline, downsampling data to 500 Hz and applying bandpass filters between 1 and 50 Hz. To prevent confounding from overt muscle movement and jaw articulation, analysis windows were restricted strictly to the three-second silent processing epochs following auditory word onset. Crucially, a board-certified epileptologist visually inspected every epoch, systematically discarding intervals containing epileptiform spikes or sharp waves to isolate pure task-induced network dynamics from transient spike disruptions. The researchers computed phase-based functional interactions between six primary regions of interest: bilateral motor cortices, left inferior frontal cortex corresponding to Broca’s area, left superior temporal cortex encompassing Wernicke’s area, and their contralateral right-hemisphere homologs.
To quantify true neural synchronization while avoiding spurious volume conduction through scalp tissues, the investigators calculated the weighted Phase Lag Index across 1,953 electrode pairs, subsequently averaging values across twelve anatomical region-to-region pairs. Analysis focused primarily on the theta frequency band, spanning four to eight hertz, given its robust mechanistic role in lexical retrieval, verbal memory, and hippocampal communication.
During both the verb generation and repetition tasks, patients displayed marked hyperconnectivity between the right inferior frontal region and bilateral motor cortices relative to controls. Analysis of core language regions revealed an unexpected finding: functional connectivity between the left inferior frontal and left superior temporal regions was significantly elevated in children with the disorder during the verb generation task. Prior resting-state functional magnetic resonance imaging literature had indicated reduced connectivity within left-hemisphere language pathways, yet task-based electrophysiological assessment uncovered that semantic retrieval demands elicited pronounced synchronization between these canonical expressive and receptive language nodes.
To assess how these abnormal network patterns relate to cognitive performance, the researchers examined associations between regional connectivity metrics and standardized phonological awareness scores. Among children with the epilepsy syndrome, high interhemispheric connectivity across opposite hemispheres during verb generation and repetition was strongly associated with poorer phonological awareness scores. Significant negative correlations emerged in five of six interhemispheric pairs during verb generation, with statistical significance surviving correction for multiple comparisons. By contrast, intra-hemispheric connectivity within the same hemisphere displayed no meaningful correlation with phonological scores, and resting-state connectivity measures in patients showed no relation to cognitive performance. In control children, higher connectivity between right motor and right temporal regions during resting correlated with lower scores, but task-based connections showed no broad impairment patterns.
Exploratory subgroup evaluations addressed the potential confounding influence of anti-seizure medication, as seventeen of the thirty-one patients received active pharmacotherapy with levetiracetam or oxcarbazepine. Stratifying patients into medicated and unmedicated cohorts revealed that elevated connectivity values were most pronounced in children receiving anti-seizure drugs, while unmedicated patients exhibited intermediate values closer to control baselines. However, standardized clinical metrics of epilepsy severity—including lifetime seizure counts, nocturnal seizure predominance, prolonged seizure duration, and secondary generalization—did not differ between medicated and unmedicated subgroups. Supplementary time-window evaluations dividing epochs into one-second intervals showed that group differences emerged primarily between 0.5 and 1.5 seconds following noun presentation, aligning with the expected temporal dynamics of lexical access and selection.
The study authors propose that task-induced hyperconnectivity may reflect delayed maturation or atypical reorganization of the developing language network. In typical neurodevelopment, children progressively transition from bilateral, diffuse cortical activation toward streamlined, left-lateralized networks that execute language tasks with minimal metabolic and network overhead. The elevated interhemispheric connectivity observed in affected children suggests an inability to segregate cortical functions, requiring the compensatory recruitment of right-hemisphere homologs and primary motor cortices to sustain task execution. Despite these profound neurophysiological differences, patient task accuracy during the EEG trials remained comparable to control levels, supporting the hypothesis that excessive connectivity represents an inefficient functional compensation that nonetheless incurs measurable cognitive costs on broader phonological processing.
Several critical methodological considerations limit the direct clinical generalization of these findings. Sensitivity analyses demonstrated that when one patient who exhibited both the highest spike burden and lowest phonological awareness score was excluded from analysis, group-level connectivity elevations remained robust, but the statistical significance of the connectivity-cognition correlation attenuated. This indicates that extreme clinical phenotypes exert substantial leverage in modest cohort sizes. Furthermore, scalp electroencephalography inherently lacks the precise anatomical source localization provided by structural magnetic resonance imaging, and the standardized vocabulary tasks do not encompass the full syntactic and communicative complexity of everyday spontaneous speech. The authors emphasize that longitudinal studies tracking pediatric cohorts through seizure remission and prospective drug trials are essential to determine whether targeted noninvasive neuromodulation can normalize these hyperconnected circuits.
Subject of Research: Functional connectivity in pediatric self-limited epilepsy with centrotemporal spikes during language tasks
Article Title: Elevated Connectivity During Language Processing Is Associated With Cognitive Performance in SeLECTS
Article References: Qi, W., Lee, K., Nix, K. C., Menchaca, M., Shé, X., Santa Maria, L., Wu, W., He, Z., & Baumer, F. M. (2026). Elevated Connectivity During Language Processing Is Associated With Cognitive Performance in SeLECTS. Annals of Clinical and Translational Neurology, 13(11), 2309-2322. https://doi.org/10.1002/acn3.70369
Image Credits: AI Generated
DOI: 10.1002/acn3.70369
Keywords: SeLECTS, pediatric epilepsy, functional connectivity, electroencephalography, language processing, phonological awareness, theta band, weighted Phase Lag Index, verb generation, inferior frontal cortex, superior temporal cortex, interhemispheric connectivity, anti-seizure medication, Broca's area, Wernicke's area, cognitive performance
News Source: Kendall Mcintyre. (November 12, 2026). Brain Hyperconnectivity During Speech Processing Links to Language Impairments in Pediatric Epilepsy. Scienmag.
