Investigating Cortical Responses to Noise-Vocoded Speech in Children with Normal Hearing Using Functional Near-Infrared Spectroscopy (fNIRS)

Abstract

Whilst functional neuroimaging has been used to investigate cortical processing of degraded speech in adults, much less is known about how these signals are processed in children. An enhanced understanding of cortical correlates of poor speech perception in children would be highly valuable to oral communication applications, including hearing devices. We utilised vocoded speech stimuli to investigate brain responses to degraded speech in 29 normally hearing children aged 6-12 years. Intelligibility of the speech stimuli was altered in two ways by (i) reducing the number of spectral channels and (ii) reducing the amplitude modulation depth of the signal. A total of five different noise-vocoded conditions (with zero, partial or high intelligibility) were presented in an event-related format whilst participants underwent functional near-infrared spectroscopy (fNIRS) neuroimaging. Participants completed a word recognition task during imaging, as well as a separate behavioural speech perception assessment. fNIRS recordings revealed statistically significant sensitivity to stimulus intelligibility across several brain regions. More intelligible stimuli elicited stronger responses in temporal regions, predominantly within the left hemisphere, while right inferior parietal regions showed an opposite, negative relationship. Although there was some evidence that partially intelligible stimuli elicited the strongest responses in the left inferior frontal cortex, a region previous studies have suggested is associated with effortful listening in adults, this effect did not reach statistical significance. These results further our understanding of cortical mechanisms underlying successful speech perception in children. Furthermore, fNIRS holds promise as a clinical technique to help assess speech intelligibility in paediatric populations.

Keywords: auditory processing; language; neuroimaging; speech perception; temporal cortex.

Fig. 1

fNIRS measurement channel locations and a priori regions of interest. The channels outlined in blue form the superior temporal ROIs, which are located symmetrically in the LH and RH. The channels outlined in orange form the left inferior frontal ROI. The channel outlined in purple forms the left posterior temporal ROI

Fig. 2

Mean speech perception scores pre- and post-imaging for the five stimulus conditions. Participants failed to identify any keywords in the two unintelligible conditions (1-ch-F and 8-ch-Z). Scores in the two partially intelligible conditions (4-chF and 8-ch-R) increased from pre- to post-imaging, suggesting a learning effect. Performance was close to ceiling level in the 8-ch-F condition. Error bars show 95 % confidence intervals corrected for a repeated-measures design following O’Brien and Cousineau (2014)

Fig. 3

Mean accuracy and response time for the word recognition task completed during fNIRS imaging. Percentage of correct responses is shown in panel a, and mean response time is shown in panel b. For both measures, performance remained consistent across fNIRS runs but differed significantly between stimulus conditions (p < 0.001). Response accuracy was approximately at chance (50 %) under the two unintelligible listing conditions (1-ch-F and 8-ch-Z) but improved as stimuli became more intelligible. Response times peaked under the partially intelligible (4-ch-F and 8-ch-R) conditions and were at an intermediate level under the most intelligible condition (8-ch-F). Error bars show 95 % confidence intervals corrected for a repeated-measures design following O’Brien and Cousineau (2014)

Fig. 4

Group-level channel-wise statistical map for the main effect of stimulus intelligibility. The map is thresholded at an uncorrected p-value of .05. The nine channels in each hemisphere that showed a significant result after FDR correction (q < .05) are highlighted. Note that the map is interpolated from single-channel results and the overlay on the cortical surface is for illustrative purposes only

Fig. 5

Group-level channel-wise relationships between fNIRS response amplitude and stimulus intelligibility. Rows a–c show the results of statistical significance testing (uncorrected p-values, thresholded at p < .05) for 0th-order, 1st-order (linear), and 2nd-order (quadratic) effects, respectively. Note that responses were averaged across conditions of similar stimulus intelligibility (i.e., across 1-ch-F and 8-ch-Z, and across 4-ch-F and 8-ch-R). Individual channels exhibiting significant effects after FDR correction (q < .05) are highlighted. Two separate colour scales are used to indicate effects in either direction, positive or negative. Note that the maps are interpolated from single-channel results and the overlay on the cortical surface is for illustrative purposes only

Fig. 6

Mean contrast values (i.e., ERAs relative to silence, arbitrary units) for a priori and post hoc ROIs. Panels a–c show bar plots derived from a priori ROIs: left auditory, right auditory, and left inferior frontal gyrus (LIFG), respectively. Panels d–f show bar plots derived from post hoc ROIs: left posterior temporal (PT), right superior temporal/inferior frontal (ST/IF), and bilateral inferior parietal/lateral temporal (IP/LT). The post hoc ROIs were defined based on significant effects in the channel-wise fNIRS analyses illustrated in Fig. 5. Thick black lines show the best quadratic fit to the data. Response amplitude increased with rising stimulus intelligibility in the left, but not right, auditory ROI. Response amplitude also increased with rising stimulus intelligibility in the left PT and right ST/IF ROIs. In the LIFG, response amplitude was highest for partially intelligent speech. In bilateral IP/LT regions, a trend towards deactivation was observed, with the greatest deactivation occurring for partially intelligible speech. Error bars show 95 % confidence intervals corrected for a repeated-measures design following O’Brien and Cousineau (2014)

Fig. 7

Event-averaged haemodynamic time courses for a priori and post hoc ROIs. The red and blue traces show estimated changes in the concentration of HbO and HbR, respectively (average response to silent trials subtracted out). The shaded area represents the 95 % confidence interval around the group mean. The first three rows show time courses for a priori ROIs: left auditory, right auditory, and left inferior frontal gyrus (LIFG), respectively. The lowermost three rows show time courses for post hoc ROIs: left posterior temporal (PT), right superior temporal/inferior frontal (ST/IF), and bilateral inferior parietal/lateral temporal (IP/LT)

Fig. 8

Mean contrast values (i.e., ERAs relative to silence, arbitrary units) as a function of stimulus intelligibility and hemisphere in a priori and post hoc ROIs. Panel a shows results for the a priori auditory ROIs (channels 29 and 33 in the LH and channels 7 and 12 in the RH). Response amplitude increases with rising stimulus intelligibility, but only in the LH, as confirmed by a significant stimulus intelligibility × hemisphere interaction (p < 0.001). Panel b shows results for the post hoc ROIs, which are the channels that showed the strongest evidence of a positive linear relationship with stimulus intelligibility (asymmetrical channels 23, 28, and 29 in the LH and channels 2, 6, and 11 in the RH). In these ROIs, responses in the LH and RH showed a more similar pattern, increasing as stimuli became more intelligible. Error bars show 95 % confidence intervals corrected for a repeated-measures design following O’Brien and Cousineau (2014)

Fig. 9

Scatter plots of behavioural performance against participant age. Positive correlations were observed between speech intelligibility and age (panel a) and between accuracy on the word recognition task conducted during fNIRS imaging and age (panel b)