The cerebral hemodynamic response to phonetic changes of speech in preterm and term infants: The impact of postmenstrual age

Abstract

Higher brain dysfunction, such as language delay, is a major concern among preterm infants. Cerebral substrates of cognitive development in preterm infants remain elusive, partly because of limited methods. The present study focuses on hemodynamic response patterns for brain function by using near-infrared spectroscopy. Specifically, the study investigates gestational differences in the hemodynamic response pattern evoked in response to phonetic changes of speech and cerebral hemispheric specialization of the auditory area in preterm infants (n = 60) and term infants (n = 20). Eighty neonates born between 26 and 41 weeks of gestational age (GA) were tested from 33 to 41 weeks of postmenstrual age (PMA). We analyzed the hemodynamic response pattern to phonemic and prosodic contrasts for multiple channels on temporal regions and the laterality index of the auditory area. Preterm infants younger than 39 weeks of PMA showed significantly atypical hemodynamic patterns, with an inverted response shape. Partial correlation analysis of the typicality score of hemodynamic response revealed a significant positive correlation with PMA. The laterality index of preterm infants from 39 weeks of PMA demonstrated a tendency rightward dominance for prosodic changes similar to term infants. We provide new evidence that alterations in hemodynamic regulation and the functional system for phonemic and prosodic processing in preterm infants catch up by their projected due dates.

Keywords: BOLD, blood oxygenation level dependent; BPD, bronchopulmonary dysplasia; Deoxy, deoxygenated; GA, gestational age; HRF, hemodynamic response function; IQR, interquartile range; Laterality; MMN, mismatch negativity; NEC, necrotizing enterocolitis; Near-infrared spectroscopy; Oxy, oxygenated; PMA, postmenstrual age; PNA, postnatal age; Preterm infants; ROI, region of interest; SOA, stimulus onset asynchrony; Speech perception; fNIRS, functional near-infrared spectroscopy.

Fig. 1

HRF patterns in the temporal regions. (a) Location of 12 channels for each hemisphere. Each channel is represented by a number. Channels 1 to 12 are located on the left hemisphere (left) and channels 13 to 24 are located on the right hemisphere (right). (b, c) Representative examples of time courses of Hb changes. The typical HRF pattern is characterized by an increase in oxy-Hb and a slight decrease in deoxy-Hb (b), while the inverted pattern is characterized by a decrease in oxy-Hb (c).

Fig. 2

The HRF pattern according to PMA. (a) The proportion of HRF patterns for each group. The three types of categories are: “typical” showing the normal HRF for two conditions, “intermediate” showing a normal HRF for one condition, and “atypical” showing an inverted HRF for two conditions. (b) HRF-typicality scores according to stimuli and PMA group. (c) Correlation between HRF-typicality scores and PMA (R = 0.30, P = 0.006).

Fig. 3

Laterality indices for phonemic and prosodic conditions according to PMA group. A positive value means left-dominance, whereas a negative value means right-dominance. * = P < 0.05.

Supplementary Fig. S1

Some examples of HRF patterns different in categorical labels (i.e. Typical, Intermediate, Atypical). Positive responses (positive HRF typicality score) for both conditions are categorized to “Typical”. Positive and negative responses for one of the two conditions are categorized to “Intermediate”. Negative responses for both of the conditions are categorized to “Atypical”.