Meta-analysis of the visuospatial aftereffects of prism adaptation, with two novel experiments

Abstract

We present a meta-analysis of the effects of visuomotor adaptation to leftward displacing prisms on visuospatial judgements in healthy people, as assessed by perceptual (landmark) and manual versions of the line bisection task. To supplement previously published datasets, we report two novel experiments: Experiment 1 (n = 12) found null effects of adaptation to 10° leftward prisms on spatial bias in the landmark task, and Experiment 2 (n = 24) found null effects of 12° leftward prisms on spatial bias in a computerised line bisection task. Including these data, we considered 17 experiments for the landmark task (total n = 256), and 12 experiments for line bisection (total n = 172), in which participants were adapted for between 7 and 20 min to prism strengths from 8 to 17°. A random-effects meta-analysis, with prism strength and exposure duration as moderators, confirmed robust rightward shifts in visuospatial judgements following leftward prism adaptation. The average standardised effect sizes (Cohen's d) were similar between tasks, increasing by around .1 per degree of prismatic displacement, and being boosted by a long (10 min +) period of prism exposure. However, the quality of evidence and precision of prediction was superior for the landmark task, with a higher signal-to-noise ratio within studies, and less heterogeneity between studies. We suggest that line bisection responses may be contaminated by sensorimotor aftereffects, and that the landmark task is a more suitable method for measuring true visuospatial aftereffects of prism adaptation. To harness these effects, we recommend that researchers should expose participants to 15° (or higher) leftward prisms for more than ten minutes, with upwards of 250 pointing movements. Power calculations should take account of heterogeneity in the true effect size between studies; and further investigation of the factors underlying this heterogeneity will help to refine optimally-effective methods.