Abstracts

Rationale: Children begin to imitate speech sounds during infancy and distinguish a wider variety of spectral-temporal features than adults. Their auditory perception gradually becomes tuned to those relevant to their daily living, including speech sounds. Investigators suggest that the anterior superior-temporal gyrus (STG) processes auditory stimuli in a sustained manner to encode the phonetic features, whereas the posterior STG detects the onset of auditory stimuli to decode the boundary between sounds. We determined if such a functional parcellation is developed and strengthened after birth, using high gamma activity modulated by sound stimuli.

Methods: We studied 2,039 non-epileptic intracranial electrode sites sampled from 32 patients with focal epilepsy (age range: 8 months to 28 years) who underwent extraoperative electrocorticography. Patients were given forward- and backward-speech sounds as well as signal-correlated noises during a task-free condition. On a spatially normalized brain surface image, we animated the group-level dynamics of high gamma activity at 70-110 Hz as a summary measure of stimulus-induced cortical activation (Figure 1A). We determined if the anterior STG shows sustained high gamma augmentation during sound stimulus presentation, whereas the posterior STG shows transient augmentation at stimulus onset. Linear regression determined the age effect on sound-related high gamma activity at each site and during each time window (Figure 1B). We determined if older individuals show more sustained high gamma augmentation in the anterior STG and more ephemeral augmentation in the posterior STG. We finally determined if speech sounds, compared to noises, induce a higher age-dependent growth of high gamma augmentation.

Results: Both speech sounds and noises induced high gamma augmentation in the STG (Figure 1A). The group-level atlas indicates that high gamma augmentation was sustained in the anterior STG 5-6 cm from the temporal lobe tip but transient in the posterior STG 8-9 cm from the tip. We found a double dissociation between the anatomical locations and age-dependent changes in high gamma dynamics (Figure 1B). The anterior STG showed an age-dependent growth of high gamma augmentation at 100-400 ms after speech sound onset (Figure 2A); noises showed a less prominent age-dependent growth of high gamma augmentation (Figures 2B). The differential age effect between stimulus types was maximal in the anterior STG 300-400 ms from stimulus onset (cohen’s d = 3.60; t-value = 35.99; 95% CI = 3.06 to 4.14; p < 0.001; Figure 2C). The posterior STG showed an age-dependent reduction of high gamma augmentation at 200-600 ms after the onset of speech sounds and noises (Figures 2A and 2B).