Listen and Learn: An Examination of Perceptual Anchoring in Individuals with Normal Hearing

<p dir="ltr">Statistical learning is defined as the ability to extract the intrinsic properties (i.e., regularities and irregularities) from sensory input over time. An important feature of auditory statistical learning is its apparently automatic nature; listeners can blithely attend to an unrelated task, ignorant of ongoing changes in the sound environment, yet their brain activity is constantly tracking changes in the sound environment and adjusting to ensure that important changes are not missed. <i>In vivo </i>(animal) experiments demonstrate a clear role for rapid neural ‘adaptation’ in this process whereby neurons track the statistical structure of sound environments, constantly altering their sensitivity (i.e., adapting) to improve listening performance (Dean et al., 2005, 2008). This learning occurs on a moment-by-moment (shortterm) basis, and can be held as a longer term ‘memory’ such that adaptation to an environment is faster if that environment has been encountered previously, a phenomenon entitled meta-adaptation. This longer-term learning effect can be disrupted when auditory neurons are dissociated from higher brain centres that provide sensory feedback through efferent neural pathways (Robinson et al., 2016). Recent studies have exploited statistical learning techniques to determine whether individuals with dyslexia are less able to adapt to and learn sounds (meta-adaptive effect; Agus et al., 2014; Daikhin et al., 2017) – an impairment named, the perceptual anchoring deficit hypothesis. An impairment that I hypothesise may reflect impaired efferent neural pathways, and thus, impaired meta-adaptive processing.</p><p dir="ltr">In this dissertation my aims were to 1) assess longer-term statistical learning (i.e., meta-adaptation) in human listeners using the perceptual anchoring paradigm; 2) investigate the contribution of frontal cortical regions in statistical learning; and 3) explore the contribution of impaired statistical learning in dyslexia. I hypothesised that meta-adaptive processes would be supported by frontal cortical regions, and that this process would be impaired or impoverished in individuals with dyslexia. I achieved these goals by investigating the learnability of broadband (i.e., white noise) stimuli that had been used in past perceptual anchoring tasks and investigated whether anchoring performance changes when using more statistically definable stimuli (i.e., pure-tone sequences) (Agus et al., 2010, 2014; Barascud et al., 2016; Bianco et al., 2020; Daikhin et al., 2017). I attempted to disrupt SL in human listeners by applying rapid transcranial magnetic stimulation (rTMS) to frontal neural regions – a region implicated in the processing of and integration of sequential information and structured tone sequences (Abla & Okanoya, 2008; Batterink et al., 2019; Fedorenko et al., 2012). Finally, I summarised, via systematic review and meta-analysis, perceptual anchoring findings to date in order to establish the strength of the anchoring deficit hypothesis.</p><p dir="ltr">Results show that listeners were able to anchor to both white-noise and pure-tone sequences, albeit with reduced variance for pure-tone sequences, and that successful anchoring was largely determined by the repeating nature of the stimulus. My attempts to impair anchoring ability by disrupting frontal cortices, via rTMS, did not significantly alter overall anchoring performance. However, it did impact how listeners performed the task over time, highlighting a potential role for frontal regions in time-course of SL. While I was unable to directly investigate anchoring ability in individuals with dyslexia due to lockdown protocols during the pandemic, our meta-analysis supports the finding that anchoring deficits are a feature of this neurodevelopmental condition and that future research should target how anchoring unfolds developmentally, and more broadly, how variations in anchoring ability may reflect a unique and underexplored aspect of hearing and listening in neurodiverse populations.</p>