Neurophysiology
ABR. Typical ABRs were observed in all neonates. Grand-average waveforms, and wave V amplitude and latency distributions are illustrated in Figure 1. Statistical analyses (Table 3) revealed that wave V amplitude was smaller in LGA compared to AGA neonates (P = .01). However, after normalization by the pre-stimulus neural activity, these differences washed out (P = .23). No group differences were observed in wave V peak latency (P = .07).
FFR. As it can be seen in Figure 2, the FFR is a neurophysiological response that reflects the neural encoding of the eliciting sound with striking faithfulness. Clear FFR waveforms were recorded in both groups, LGA and AGA, although in the LGA group a notably smaller amplitude through the entire FFR duration was observed compared to the AGA group (Figure 2B). Notice, however, that the background EEG was similar in the two groups (no group differences in pre-stimulus RMS: P = .75; Figure 3A; Table 3). Neural lag was also comparable across groups (P = .10; Figure 3B; Table 3).
To quantify the FFR, the F0 spectral amplitude and its normalization (SNR) were retrieved by means of the FFT for the responses to both the stimulus consonant transition and the vowel regions in each group (Figure 2C; Figure 3C to F). Smaller spectral amplitudes were found in LGA compared to the AGA group for the consonant transition (P = .002) and the vowel regions (P = .008). Similar results were obtained after normalization (i.e., after retrieving the FFR’s SNR), showing that the LGA group had a smaller SNR than the AGA group in both regions (consonant transition region: P = .01; vowel region:P = .005).
To further characterize our data, we tested whether our results could be explained by the neonates’ adipose tissue. For that, a matrix of correlations between neonatal BMI and all FFR parameters was computed across the entire final sample (N = 50, see Figure 4). Strong significant correlations between neonatal BMI and spectral amplitude for the consonant transition (r = -0.54, P <.001) and for the vowel region (r = -0.57, P <.001) were found, as well as with the normalized spectral amplitude (SNR) for the vowel region (ρ = -0.42, P = .004) (Figure 4). In line with previous literature in which a significant relation was observed between maternal pre-gestational BMI and/or gestational weight gain (GWG) and offspring cognitive outcomes66,67, we carried out an identical statistical approach replacing the neonatal variable with the maternal variables. These analyses yielded no significant correlations.