Measuring the capacity for auditory system plasticity: An examination of performance gains during initial exposure to auditory-targeted cognitive training in schizophrenia

Abstract

Auditory-Targeted Cognitive Training (ATCT), which aims to improve auditory information processing efficiency, has shown great promise for remediating cognitive deficits in schizophrenia (SZ). However, there is substantial heterogeneity in the degree of cognitive gains made during ATCT, and some patients show negligible benefit after completing therapeutic doses of training. Identifying individual differences that can be measured early in the course of ATCT and that predict subsequent cognitive benefits from the intervention is therefore important. The present study calculated a variety of performance metrics during the initial hour of exposure to ATCT Sound Sweeps, a frequency discrimination time-order judgment task, and investigated the relationships of these metrics to demographic, clinical, and cognitive characteristics of SZ patients.

Thirty-seven SZ outpatients completed measures of auditory attention, working memory, verbal memory, and executive functioning, followed by 1 h of Sound Sweeps training. Performance metrics, calculated after the first training level, the first training stage (Levels 1–4), and the entire hour of training included baseline and best auditory processing speed (APS) scores, as well as percent improvement in APS after training. The number of training levels completed by each participant was also calculated.

Baseline and best APS correlated with performance in all cognitive domains, whereas APS improvements only correlated with verbal memory. Number of training levels completed was marginally associated with auditory attention only.

Conclusions: Sound Sweeps performance correlates with a range of neurocognitive abilities. APS improvement may provide a particularly sensitive index of “plasticity potential” within the neural network underlying verbal learning and memory.

Introduction

Neurocognitive deficits represent a core feature of schizophrenia (SZ) that impinge upon daily psychosocial functioning (Green, 1996, Green et al., 2000), and efforts at remediating cognitive deficits have generally shown a modest degree of efficacy at the group level (McGurk et al., 2007, Wykes et al., 2011). As is often the case for psychiatric interventions, cognitive remediation modalities are typically developed for the average patient and implemented in the absence of knowledge about individual variation in genes, brain function, pathophysiology, and environment that might influence treatment outcomes. This one-size-fits-all approach to cognitive remediation is particularly problematic given data that suggest that up to 45% of people with SZ demonstrate virtually no cognitive enhancement after undergoing a therapeutic dose (≥ 32 h) of computerized cognitive training (Murthy et al., 2012). For patients and clinicians, the costs associated with these time- and resource-intensive interventions can be prohibitive. Thus, cognitive training is an excellent example of a treatment that may benefit from a recently announced “precision medicine” initiative by NIH, which aims to promote the systematic investigation of individual differences that play a role in illness and health. Ultimately, the initiative aims to facilitate data-driven prediction of benefit for individual patients from specific treatments at any point during the course of illness. Put simply, precision medicine strives to provide the “right treatment” to the “right person” at the “right time.”

Despite the substantial advances made in cognitive remediation over recent years, the ability to characterize the SZ patients for whom any form of cognitive remediation is the “right” intervention continues to elude practitioners. Accordingly, the present study aims to investigate several auditory processing measures for their ability to reflect early neural target engagement during initial exposure to Auditory Targeted Cognitive Training (ATCT), a computerized intervention that has shown particular promise for enhancing cognition in SZ (Fisher et al., 2009, Fisher et al., 2015, Popov et al., 2011).

While conventional cognitive remediation techniques typically target cognition from the “top-down” (e.g. teaching memory encoding strategies, and problem solving approaches), ATCT focuses on “bottom-up” or feed-forward training of auditory processing fidelity and efficiency, while simultaneously harnessing attention and working memory operations. ATCT explicitly employs known mechanisms to maximize cortical neuroplasticity, for example, by delivering exercises with specifically-defined learning targets delivered at high intensity (greater than 1000 trials throughout a full course of training), and by maintaining difficulty levels that are carefully titrated in accordance with individual patient performance (Merzenich et al., 2013). Moreover, correct responses are reinforced with sounds and visual animations, consistent with literature suggesting a neuromodulatory effect of subcortical reward processing centers on cortical representations of selectively attended sensory inputs (Merzenich et al., 2014, Vinogradov et al., 2012). ATCT thus aims to efficiently modify the frontotemporal cortical dynamics subserving both basic perceptual processes and higher-order cognitive operations (Vinogradov et al., 2012).

Studies examining the efficacy of ATCT in SZ patients have shown large improvements (d = 0.86–0.89) in verbal learning and memory, verbal working memory, and global cognition after 40–50 h of training; moderate improvements have also been detected in non-trained visual problems solving skills (Fisher et al., 2009, Fisher et al., 2015). Furthermore, in a recent multi-site study, significant gains in MATRICS Consensus Cognitive Battery (MCCB) composite and verbal learning scores were observed after 20 h of training. While these gains no longer achieved statistical significance after 40 h of training, perhaps due in part to subject attrition, the effect sizes (d ≈ 0.39) remained non-trivial (Keefe et al., 2012).

Despite its apparent efficacy at the group level, individual response to ATCT is highly variable (Murthy et al., 2012). Although some patient characteristics, such as self-reported anticipatory pleasure, appear to correlate with ATCT response (Fisher et al., 2015), there are currently no established methods for identifying early in treatment (e.g. within the first hour) the individuals most likely to benefit from ATCT. Auditory perceptual improvements (i.e. auditory “tuning”) gained during treatment may be a key predictor of ATCT response, as they have been shown to correspond to overall degree of cognitive enhancement (Murthy et al., 2012, Popov et al., 2011, Popov et al., 2012, Popov et al., 2015). Notably, Fisher et al. (2015) found that auditory processing speed improvements after 20 h of training correlated with degree of cognitive enhancement after up to 40 h of training, suggesting that these early auditory processing improvements may reflect the “plasticity potential” of the frontotemporal network and thus index the likelihood of cognitive benefit from ATCT. Since the most dramatic improvements in auditory processing have been shown to occur very early in the course of training, with maximal gains evident after the first training session and incremental gains following subsequent training sessions (Menning et al., 2000), detailed examination of the auditory perceptual dynamics occurring within the initial exposure to ATCT might thus account for some variation in individual training response and subsequently inform future predictive algorithms for guiding treatment of SZ.

Challenges exist, however, in quantifying the perceptual gains made within or across training sessions or patients, given the individualized and continually adaptive nature of ATCT, an intervention that was primarily designed for clinical rather than academic purposes. As such, the aim of the present study was to examine several measures of auditory perceptual improvement during the initial hour of ATCT for their relationships to critical demographic, clinical, and cognitive characteristics of SZ patients. Several such performance metrics have been utilized in previous studies, including percentage of auditory frequency discrimination exercises completed (i.e. the percentage of training trials completed out of the total number of trials available; Fisher et al., 2009), and auditory processing speed (APS; i.e. the length of auditory stimuli for which participants are able to make accurate time-order judgments; Fisher et al., 2015, Keefe et al., 2012, Mahncke et al., 2006, Murthy et al., 2012, Smith et al., 2009). While relationships between these auditory processing measures and ATCT-related cognitive gains have been detected across therapeutic training intervals (e.g., 20–40 h; Fisher et al., 2009, Fisher et al., 2015), no previous studies have examined whether auditory processing improvements during the initial exposure to ATCT are related to demographic, clinical, and cognitive characteristics of patients at baseline. We hypothesized that better auditory processing at baseline, as well as larger improvements in auditory processing after 1 h of ATCT, would be associated with 1) younger patient age, 2) later age of illness onset, 3) less severe clinical symptoms, and 4) better cognitive performance.