Reliability of VPA Learning Over Repeat Exposures Across an Extended Interval Using Automated Remote Method

Abstract

AbstractBackgroundThe current study examined the test‐retest reliability of a newly developed Learning Over Repeat Exposures (LORE) paradigm using Verbal Paired‐Associates. This LORE paradigm involves the acquisition of a set of 8 verbal paired associates over a 5‐day “burst”. In this investigation, two bursts, separated by a nine‐month interval, were compared to assess the stability of novel learning outcome measures, such as the learning slope (between day 1 and the day of max performance) and first overnight forgetting.MethodAn opportunistically recruited sample of 14 participants (out of an initial pilot study of N = 20) completed the VPA‐LORE task using the Cambridge Cognition ‘NeuroVocalix’ platform, which employs text‐to‐speech and automated speech recognition technology to remotely administer and score the task, enabling its potential use at scale. The task structure consists of an immediate recall following presentation on the first day of a burst, and a single recall attempt followed by a presentation on subsequent days. As with the pilot study, neither floor nor ceiling effects were observed, and performance was in the expected range.ResultThe observed moderate to strong correlations between the two bursts revealed good reliability for the measures examined: learning slope from day 1 to day of max score, r s = .815, first day immediate recall, r s = .736, max pairs learned, r s = .707, and day 1 forgetting r s = .702. In total, participants missed only a few visits, with six missing sessions out of a total of 140 sessions, representing 4.3% of all sessions. The retention of 14 participants from the original pilot study also highlights the feasibility and acceptability of this method.ConclusionThese results suggest that the VPA‐LORE paradigm, delivered via the NeuroVocalix platform, is a stable and reliable assessment tool even when considering extended intervals between bursts and accounting for missing data. The novel learning measures derived from the VPA‐LORE paradigm provide valuable insights into learning processes and could further enhance cognitive assessment. Future research could explore the applicability of the paradigm for clinical populations and longitudinal studies in cognitive assessment.