Skip to main content

The listening challenge: How ageing brains adapt

Periodic Reporting for period 3 - AUDADAPT (The listening challenge: How ageing brains adapt)

Reporting period: 2019-01-01 to 2020-06-30

Humans in principle adapt well to sensory degradations. In order to do so, our cognitive strategies need to adjust accordingly (a process we term “adaptive control”). The auditory sensory modality poses an excellent, although under-utilised, research model to understand these adjustments, their neural basis, and their large variation amongst individuals. Hearing abilities begin to decline already in the fourth life decade, and our guiding hypothesis is that individuals differ in the extent to which they are neurally, cognitively, and psychologically equipped to adapt to this sensory decline.

Our project pursues three specific aims: (1) We will first specify the neural dynamics of “adaptive control” in the under-studied target group of middle-aged listeners compared to young listeners. We will employ advanced multi-modal neuroimaging (EEG and fMRI) markers and a flexible experimental design of listening challenges. (2) Based on the parameters established in (1), we will explain inter-individual differences in adaptive control in a large-scale sample of middle-aged listeners, and aim to re-test each individual again after approximately two years. These data will lead to (3) where we will employ statistical models that incorporate a broader context of audiological, cognitive skill, and personality markers and reconstructs longitudinal “trajectories of change” in adaptive control over the middle-age life span.

Pursuing these aims will help establish a new theoretical framework for the adaptive ageing brain. The project will further break new ground for future classification and treatment of hearing difficulties, and for developing individualised hearing solutions. This project thus bears the potential to challenge and to transform current understanding and concepts of the ageing human individual.
We began to validate a set of neural markers relevant for the processing of speech under challenging conditions. Building on previous results, we focused on two key neurophysiological measures: the neural tracking of speech by slow neural oscillations (1–7 Hz) in auditory cortex , and more domain-general dynamic modulation of alpha oscillations (8–13 Hz) often associated with functional inhibition in control of attention.
We built on and extended existing experimental paradigms known to reliably tap into selective auditory processing. We decided on a battery of cognitive tests, personality profiling and audiometric measurement that would complement EEG (and fMRI) measurements that would allow for extended modelling of neural and non-neural markers of adaptive control in WP2 and 3.

We recruited an initial sample of ~40 healthy, middle-aged adults who underwent sessions of behavioral and EEG recording. A first set of multivariate analyses revealed a separation of neural and behavioral variables independent of age (Tune et al., 2018, Eur J Neurosci). A tentative conclusion to guide the large-sample analyses of WP2 data is that alpha power lateralization is neither a sufficient nor necessary neural strategy for an individual’s auditory spatial attention, and that we need to further investigate the use of alternative compensatory mechanisms.

We investigated graph-theoretical brain network models derived from MEG and fMRI, and entropy of EEG signals, in relation to auditory processing. (Alavash et al. 2017, Network Neurosci; APAN/SfN, 2016). Also, the overall level of neural irregularity increased with age, paralleled by a decrease in variability over time, which likely indexes senescent change (e.g. changes in excitation/inhibition balance). Second, states of increased irregularity prior to stimulus onset led to optimized sensory encoding and influenced ensuing perceptual decisions regardless of age (Waschke, Wöstmann & Obleser, Sci Rep 2017).

In parallel, we also developed of a novel experimental paradigm to simultaneously investigate, the selective tracking of speech as well as the dynamics of lateralized alpha power at the single trial level. We extended our previously established dichotic listening tasks two different listening cues that induced spatial auditory attention and semantic predictions, respectively. This paradigm allowed us to investigate more directly the impact of additional top-down factors on listening behavior, and how ageing individuals would differ in their ability to utilize these cues to guide speech comprehension.

We continued investigating network dynamics in fMRI data and operationalized the functional implementation of adaptive control in terms of reconfiguration of resting state brain networks in adaptation to a listening challenge. Using our previously validated network modelling approach (Alavash et al., 2017 Network Neuroscience; Alavash et al., 2018 NeuroImage), we were able to explain the inter-individual variability in listening success by modular reconfiguration of resting state brain networks in adaptation to the listening task. The key finding of this study was that higher network modularity of an auditory-control brain network relative to its baseline predicted individuals’ listening success (Alavash et al., in prep.; SAMBA, 2018; RSBC, 2018).
Where are we / where have we been moving beyond the state of the art with this project?
1. PUSHING THE ENVELOPE OF AUDITION IN MIDDLE-AGE ADULTHOOD: Our work is characterized by a wholesome view on audition: It does entail audiological testing on one end, and it does entail a large-scale, whole-brain characterization of even resting state characteristics and brain dynamics on the other end. Crucially, however, this project is amongst the first to model the sensory and cognitive processes in middle-aged adult, prior to or in the absence of major hearing problems, as a process that entails all these variables. Our multi-modal neuroimaging approach that not only uses EEG and fMRI but looks at various, very different measures in each of these,

2. FUSING PSYCHOLOGICAL AND NEUROSCIENTIFIC MODELLING OF AUDITION IN MIDDLE-AGE ADULTHOOD: As part of a still ongoing examination that also marks the seamless transition from ongoing WP2 into WP3, we have started to model the neural results along with the data on cognitive fitness, personality structure, as well as objective and subjective measures of hearing acuity in a multivariate, structural equation modelling framework. This modelling work will allow us to better understand the interplay of neural and psychological mechanisms which determine how well an ageing listener adapts to sensory decline. In the coming months, we hope to further elucidate the most important factors that underlie the considerable degree of inter-individual variation in the ability to adapt to challenging listening situations.

3. PROVIDING A LONGITUDINAL PERSPECTIVE. Longitudinal data are of the essence in any form of life-span research. Any attempt at causal inference is greatly facilitated if temporal precedence and observed changes can be substantiated by repeated measures over time in the same cohort. Thus, We deem the ERC AUDADAPT cohort, which we have started to establish (currently approx. N=120 listeners aged 40 to 80) an important contribution to the field. All data will be made publicly available. As the project is still under way, the cohort is still growing and will be re-tested in the years to come using behavior, audiometry, brain stem measures, cortical EEG and fMRI plus a host of cognitive, lifestyle, and personality profiles.