Amplitude Modulations Of Speech Research Articles

Contribution of positive affect in infant directed speech: what do amplitude modulations patterns suggest?

Speech perception relies heavily on cortical entrainment of amplitude modulations in speech, which occur at different rates. Following a study by Leong et al. (2017) showing slower modulations have higher power than faster modulations in infant directed speech (IDS), we hypothesized that positive emotions in IDS might drive this pattern. Using the same analyses (spectral amplitude modulation phase hierarchy method and phase synchronization index (PSI) (Leong and Goswami, 2015)), we compared the power of isolated modulation rates (synchronous with neural oscillations) and the synchrony between them in IDS and adult directed speech (ADS), using English stimuli from Many Babies Consortium (Frank et al., 2020). We repeated the same analyses comparing happy and neutral ADS using stimuli of four native English speakers (Pell et al., 2009). Our analysis did not uncover significant power differences between IDS and ADS. However, it revealed happy ADS has higher power at slower rates, with the reversed pattern for neutral ADS (p < 0.001). PSI was higher for two faster rates in neutral ADS (p < 0.0001), as reported by Leong et al. comparing IDS versus ADS. These findings reveal novel acoustic features of vocal emotions that might be important in attracting infant attention to IDS.

Temporal Auditory Coding Features for Causal Speech Enhancement

Perceptually motivated audio signal processing and feature extraction have played a key role in the determination of high-level semantic processes and the development of emerging systems and applications, such as mobile phone telecommunication and hearing aids. In the era of deep learning, speech enhancement methods based on neural networks have seen great success, mainly operating on the log-power spectra. Although these approaches surpass the need for exhaustive feature extraction and selection, it is still unclear whether they target the important sound characteristics related to speech perception. In this study, we propose a novel set of auditory-motivated features for single-channel speech enhancement by fusing temporal envelope and temporal fine structure information in the context of vocoder-like processing. A causal gated recurrent unit (GRU) neural network is employed to recover the low-frequency amplitude modulations of speech. Experimental results indicate that the exploited system achieves considerable gains for normal-hearing and hearing-impaired listeners, in terms of objective intelligibility and quality metrics. The proposed auditory-motivated feature set achieved better objective intelligibility results compared to the conventional log-magnitude spectrogram features, while mixed results were observed for simulated listeners with hearing loss. Finally, we demonstrate that the proposed analysis/synthesis framework provides satisfactory reconstruction accuracy of speech signals.

Enhanced amplitude modulations contribute to the Lombard intelligibility benefit: Evidence from the Nijmegen Corpus of Lombard Speech.

Speakers adjust their voice when talking in noise, which is known as Lombard speech. These acoustic adjustments facilitate speech comprehension in noise relative to plain speech (i.e., speech produced in quiet). However, exactly which characteristics of Lombard speech drive this intelligibility benefit in noise remains unclear. This study assessed the contribution of enhanced amplitude modulations to the Lombard speech intelligibility benefit by demonstrating that (1) native speakers of Dutch in the Nijmegen Corpus of Lombard Speech produce more pronounced amplitude modulations in noise vs in quiet; (2) more enhanced amplitude modulations correlate positively with intelligibility in a speech-in-noise perception experiment; (3) transplanting the amplitude modulations from Lombard speech onto plain speech leads to an intelligibility improvement, suggesting that enhanced amplitude modulations in Lombard speech contribute towards intelligibility in noise. Results are discussed in light of recent neurobiological models of speech perception with reference to neural oscillators phase-locking to the amplitude modulations in speech, guiding the processing of speech.

Atypical neural processing of rise time by adults with dyslexia

In recent studies phonological deficits in dyslexia are related to a deficit in the synchronization of neural oscillations to the dynamics of the speech envelope. The temporal features of both amplitude modulations and rise times characterize the speech envelope. Previous studies uncovered the inefficiency of the dyslexic brain to follow different amplitude modulations in speech. However, it remains to be investigated how the envelope's rise time mediates this neural processing. In this study we examined neural synchronization in students with and without dyslexia using auditory steady-state responses at theta, alpha, beta and low-gamma range oscillations (i.e., 4, 10, 20 and 40 Hz) to stimuli with different envelope rise times. Our results revealed reduced neural synchronization in the alpha, beta and low-gamma frequency ranges in dyslexia. Moreover, atypical neural synchronization was modulated by rise time for alpha and beta oscillations, showing that deficits found at 10 and 20 Hz were only evident when the envelope's rise time was significantly shortened. This impaired tracking of rise time cues may very well lead to the speech and phonological processing difficulties observed in dyslexia.

The Temporal Modulation Structure of Infant-Directed Speech

The temporal modulation structure of adult-directed speech (ADS) is thought to be encoded by neuronal oscillations in the auditory cortex that fluctuate at different temporal rates. Oscillatory activity is thought to phase-align to amplitude modulations in speech at corresponding rates, thereby supporting parsing of the signal into linguistically relevant units. The temporal modulation structure of infant-directed speech (IDS) is unexplored. Here we compare the amplitude modulation (AM) structure of IDS recorded from mothers speaking, over three occasions, to their 7-, 9-, and 11-month-old infants, and the same mothers speaking ADS. Analysis of the modulation spectrum in each case revealed that modulation energy in the theta band was significantly greater in ADS than in IDS, whereas in the delta band, modulation energy was significantly greater for IDS than ADS. Furthermore, phase alignment between delta- and theta-band AMs was stronger in IDS compared to ADS. This remained the case when IDS and ADS were rate-normalized to control for differences in speech rate. These data indicate stronger rhythmic synchronization and acoustic temporal regularity in IDS compared to ADS, structural acoustic differences that may be important for early language learning.