Severe respiratory disease caused by human respiratory syncytial virus impairs language learning during early infancy

Abstract

Human respiratory syncytial virus infection is a leading cause of pediatric morbidity and mortality. A previous murine study showed that during severe acute respiratory infections the virus invades the central nervous system, and that infected animals evolve with long-lasting learning difficulties associated with long-term potentiation impairment in their hippocampus. We hypothesized here that human infants who presented a severe episode of respiratory syncytial virus infection before 6 months of age would develop long-term learning difficulties. We measured the acquisition of the native phoneme repertoire during the first year, a milestone in early human development, comprising a reduction in the sensitivity to the irrelevant nonnative phonetic information and an increase in the sensitivity to the information relevant for the native one. We found that infants with a history of severe respiratory infection by the human respiratory syncytial virus presented poor distinction of native and nonnative phonetic contrasts at 6 months of age, and remained atypically sensitive to nonnative contrasts at 12 months, which associated with weak communicative abilities. Our results uncover previously unknown long-term language learning difficulties associated with a single episode of severe respiratory infection by the human respiratory syncytial virus, which could relate to memory impairments.

Figure 1

The native and nonnative phonetic distinction. In (a) we illustrate the synthetic consonant–vowel stimuli previously used to study categorical perception in English, Hindi, and Spanish-speakers. This continuum comprises 16 steps (i.e. syllables) equidistant along the voiced place-of-articulation dimension from the bilabial /b/ to the dental /d/ and to the retroflex /D/, associated with the vowel /a/. Along the continuum, adult native English-speakers perceive two phonetic categories (Step1–Step6 as /ba/ and the following as /da/) while adult native Hindi-speakers perceive three (Step1–Step6 as /ba/, Step7–Step10 as /da/, and Step11–Step16 as retroflex /Da/). Adult Spanish-speakers also perceive two categories, however from Step1 to Step6 are perceived as /pa/ and the next as /ta/, due to a shorter voice onset time for these syllables in Spanish. In grey and orange, we indicate the syllables perceived as the native /pa/ and /ta/ respectively by Spanish-speaking adults, and in blue, those perceived as the nonnative /Da/ by Hindi-speaking adults. The phonetic boundaries are indicated by black vertical dotted lines. The arrow below the continuum indicates the stimulus we compared in this study to evaluate the response to phonetic and acoustic changes at the native and nonnative boundaries. In (b) we illustrate the structure of the trials. Each trial comprised the auditory presentation of four 275 ms long syllables, one every 600 ms separated by silence. The first 3 syllables were always identical, to induce habituation, while the 4th syllable remained the same in standard trials, changed to a syllable from a different phonetic category in phonetic trials or changed to a syllable of the same category in acoustic trials. Because any change between phonemes conveys a change at the acoustic and linguistic level, the subtraction of the brain response for standard and acoustic trials to the response for phonetic ones, at the native and nonnative phonetic boundaries, allowed us to quantify the purely linguistic component of the MMR.

Figure 2

The spatiotemporal clusters for the analysis of the phonetic MMR was similar at 6 and 12 months of age. By applying the cluster-based permutation technique we identified where (i.e. the electrodes) and when (i.e. the time window) the brain response to the syllable 4 differed between the phonetic and standard trials, computed across all infants, regardless of the group (hRSV and control) and type of phonetic contrast (native and nonnative), at 6 and 12 months of age. The t-maps plotted in (a) and (d) show the group of the frontal electrodes where the standard and phonetic trials differed at syllable 4, at 6 and 12 months of age, respectively, and in (b) and (e) we depict the t-values indicating the time windows when those differences remained significant, at p < 0.05, FDR corrected. In (c) and (f) we plot the time course of the grand average of the voltage amplitude for the entire standard and phonetic trials, computed over the spatiotemporal cluster corresponding to age. The soundwave images indicate the occurrence of each one of the 4 syllables of the trials. The green rectangle highlights the time window when the brain response to the standard and phonetic trials differed.

Figure 3

The V6 and V12 groups showed atypical phonetic MMR and fewer linguistic skills than the control groups. In (a) we depict the boxplots of the phonetic MMR for the native and nonnative contrasts, at syllable 3 and 4 per group. Red horizontal lines and blue diamonds correspond to the median and the mean of each group, respectively. Error bars indicate 95% confidence interval. The asterisks adjacent to the boxplots indicate significant differences from zero at p < 0.05, while the asterisks below the braces indicate significant differences between groups at the extremes of the braces. In (b) we plot the boxplots for the difference in amplitude between phonetic and acoustic trials, to extract the linguistic component of the phonetic MMR. In (c), we depict the t-map for the electrodes and t-values for the time window at the frontotemporal (black line) and parieto-occipital electrodes (grey line), over which the brain response for the syllable 1 differed from the silent baseline at 6 and 12 months of age. In (d) we plot the time course of the grand average for the syllable 1 per group and age, computed over the spatiotemporal windows at the frontal electrodes. We illustrate in (e), the mean of the first peak’s latency and in (f), the mean of the second peak polarity per group. We depict in (g) the difference in amplitude of the mean phonetic MMR between 6 and 12 months of age, in a subsample of infants evaluated at both ages, nn = nonnative; na = native. In (h) we depict the sum of words and gests understood and used by infants of the C12 and V12 groups, reported by their parents. In (i) and (j) we plot the correlations of the mean sum of words and gests handled by the infants against the phonetic MMR amplitude for the nonnative and native phonetic MMR, respectively, in C12 (red circles) and V12 groups (green circles). Pearson’s correlation coefficient and p-value are indicated for each plot.