2019-21 EC/RGC: Cortical mechanisms of temporal predictions in processing rhythmic sounds and speech

Grant title: Cortical mechanisms of temporal predictions in processing rhythmic sounds and speech

Principal (co-)investigators: Jan Schnupp, David Poeppel (MPI Frankfurt / NYU), Ryszard Auksztulewicz

Grant scheme: European Commission (EC) / Research Grants Council (RGC) Collaboration Scheme 2017/18G

Funding amount: HK$ 1,244,727

Grant duration: 1 September 2019 - 31 August 2021

Abstract: Processing complex, structured sounds such as speech and music is a challenging task for the brain. Such sounds tend to be rhythmic, and the brain is thought to be able to exploit these rhythms to generate time-specific expectations which help separate these sounds from backgrounds and ensure speedy, accurate recognition. Magneto- and electroencephalogram (MEG / EEG) recordings on human volunteers suggest that oscillations of brain activity (neural rhythms) tend to synchronize, or “entrain”, to the rhythm of the sounds we hear, which may be fundamental to speech processing (Giraud & Poeppel, 2012). Brain rhythms have also been shown to entrain to rhythmic visual or touch stimuli. It has been proposed that this rhythmic entrainment may be a manifestation of temporal expectation signals in the brain, creating windows of increased sensitivity to relevant stimuli (Morillon et al., 2015), regardless of sensory modality. Since neural entrainment has predominantly been studied using non-invasive EEG, how this entrainment alters the encoding of information at the level of groups of individual neurons  remains largely unknown. Investigating this requires detailed intracranial recordings which are only possible in animal experiments. Entrained neural oscillations may arise from local dynamics of neural firing in auditory cortex (Womelsdorf et al., 2014), or interactions between the auditory cortex and the thalamus (Musacchia et al., 2014), or the motor regions (Morillon et al., 2015). Understanding the mechanisms of entrainment not only provides deep insights into how our brains generate anticipation and temporal attention; it also has clinical relevance, as has been shown in individuals with phonological difficulties in speech processing (Power et al., 2013). To work towards this important goal, we propose an integrated research programme combining state-of-the-art animal electrophysiology, human neuroimaging, and computational modelling, which will allow us to test our hypothesis that oscillations in brain activity induce rhythmic changes in the sensitivity and frequency selectivity of individual neurons in early auditory cortex, which can be flexibly tuned to the rhythm of the acoustic inputs. Our aims are to (1) establish an animal model which, in parallel to work on human volunteers, will allow us to study the mechanisms of neural entrainment to complex sounds and its role in predictive coding in unprecedented detail; (2) disambiguate between proposed mechanisms of entrainment by studying interactions between auditory cortex and other cortical and subcortical regions. To tackle these aims, we will perform three studies, each comprising matched experiments on rats and humans.