Deconstructing spatial working memory and attention deficits in multiple sclerosis

Abstract

Objective: To investigate whether spatial working memory (WM) is impaired in multiple sclerosis (MS), and, if it is, to localize impairment to specific cognitive subprocess(es).

Method: In Experiment 1, MS and control participants performed computerized memory-span and visuomotor tasks. WM subprocesses were taxed by manipulating (1) the requirement to remember serial order, (2) delay duration, and (3) the presence of irrelevant stimuli during target presentation. In Experiment 2, recall and recognition tests varied the difficulty of WM retrieval. In Experiment 3, an attention-cueing task tested the ability to voluntarily and rapidly reorient attention.

Results: Performance was worse for MS than for control participants in both spatial recall (Exp. 1 span: 95% CIMS = [5.11, 5.57], 95% CIControls = [5.58, 6.03], p = .003, 1-tailed; Exp. 2 span: 95% CIMS = [4.44, 5.54], 95% CIControls = [5.47, 6.57], p = .006, 1-tailed) and recognition (accuracy: 95% CIMS = [0.71, 0.81], 95% CIControls = [0.79, 0.88], p = .01, 1-tailed) tests. However, there was no evidence for deficits in spatiotemporal binding, maintenance, retrieval, distractor suppression, or visuomotor processing. In contrast, MS participants were abnormally slow to reorient attention (cueing effect (ms): 95% CIMS: [90, 169], 95% CIControls: [29, 107], p = .015, 1-tailed).

Conclusions: Results suggest that, whereas spatial WM is impaired in MS, once spatial information has been adequately encoded into WM, individuals with MS are, on average, able to maintain and retrieve this information. Impoverished encoding of spatial information, however, may be due to inefficient voluntary orienting of attention.

Figure 1

Schematic illustration of the basic spatial (upper panel) and verbal (lower panel) task structure in Experiment 1. Following presentation of a target sequence, a tone sounded to indicate that participants should report the target items, either by touching the squares (in spatial-task blocks) or by pressing keyboard number keys (in verbal-task blocks). In this figure, spatial targets are indicated by black borders; in the experiment, spatial targets were indicated by a change in color. Target sequence length was adjusted following each trial by an adaptive staircase procedure. Three manipulations were conducted in a blocked fashion: (1) participants either were instructed to reproduce the serial order of the target items or were allowed to report the targets in any order, (2) the delay between the end of the target sequence and the recall cue was either short (0.25 s) or long (8.25 s), and (3) irrelevant auditory stimuli (noise bursts in the spatial task and words in the verbal task) time-locked to target onsets were either present or absent. See text for additional details.

Figure 2

Experiment 1: mean (±SE) span as a function of participant group and baseline task (i.e., participants were required to reproduce the serial order of target items, there was a minimal delay period prior to the recall cue, and no distractors were presented).

Figure 3

Experiment 1: mean (±SE) span as a function of participant group and serial order requirements for the verbal task (Panel a) and spatial task (Panel b).

Figure 4

Experiment 1: mean (±SE) span as a function of participant group and delay period for the verbal task (Panel a) and spatial task (Panel b).

Figure 5

Experiment 1: mean (±SE) span as a function of participant group and distractor presentation for the verbal task (Panel a) and spatial task (Panel b).

Figure 6

Experiment 2: Plotted in Panel a is mean (±SE) span as a function of participant group and recall task. Plotted in Panel b is mean (±SE) proportion correct as a function of participant group and recognition task.

Figure 7

Experiment 2: mean (±SE) proportion of responses that were correct rejections as a function of participant group and type of nonmatching test sequence.

Figure 8

Experiment 3: cueing paradigm. Panel a depicts an example of a valid-cue trial, where a target letter (F or T), flanked by distractors, appears at the location cued by the rectangle. Panel b depicts an example of an invalid-cue trial, where a target letter appears at the location opposite that cued by the rectangle. 80% of trials contained valid cues, and 20% contained invalid cues.

Figure 9

Experiment 3: Plotted in Panel a is mean (±SE) RT as a function of participant group, cue validity, and SOA. Plotted in Panel b is mean (±SE) cueing effect (a within-participant measure of mean RT for invalid-cue trials minus mean RT for valid-cue trials) as a function of participant group and SOA.