Scientific Papers

The effects of visual skills training on cognitive and executive functions in stroke patients: a systematic review with meta-analysis | Journal of NeuroEngineering and Rehabilitation

Description of Image

Selection of studies

The search strategy and selection process for visual skills interventions are illustrated in Fig. 1. A total of 2,343 articles on visual skills interventions were identified for further analysis in the PubMed, Medline, EMBASE, Cochrane Library, APA PsycINFO, and Web of Science databases and supplemented by 13 additional articles derived from a manual search of the reference lists. After removing duplicates, excluding articles based on title or abstract, as well as reviews, case studies, and experimental study designs, 83 articles remained for full-text consideration. 36 articles were excluded that did not specifically focus their intervention on visual skills. Fourteen studies did not examine outcome measures relevant to the present systematic review and meta-analysis, seven did not specify only stroke patients as the population, and two did not provide conclusive data.

Fig. 1
figure 1

PRISMA flow diagram illustrating the systematic literature search, study selection, and respective reasons for exclusion of records

Study characteristics

The characteristics of the included studies are listed in Table 3 and illustrate the authors, year of publication, participant characteristics, time since stroke, intervention and control groups, details of the interventions implemented, test procedures and outcome measures, the results of each group, and the methodological study quality.

Table 3 Included studies examining the effects of visual skills training on cognitive functions and activities of daily living in stroke patients

Participant characteristics

A total of 889 participants were examined in the included studies of this meta-analysis. All participants had a history of diagnosed stroke and were aged between 30 and 77 years. One study examined participants with a mean age of 30.9–39.8 years [56], 14 studies of 52.5–62.6 years [55, 58, 59, 76, 78, 80, 82,83,84, 87, 88, 90, 92, 94], six of 64.0–67.6 years [57, 77, 81, 85, 86, 91], and three of 70.7–77.7 years [79, 89, 93]. A total of 357 women and 532 men were studied. One study considered one woman and 28 men [92] and another five women and 25 men [76]. All other studies had a more balanced female-to-male ratio. The localisation was indicated by five studies with right hemisphere [58, 78, 81, 86, 94], six with various [55, 57, 59, 77, 87, 90], while all others did not provide any information. The main symptoms of the participants indicated eleven studies with cognitive dysfunction [55,56,57, 59, 78, 79, 82, 83, 86,87,88], five with motor dysfunction [58, 77, 90, 92, 94], three with visual disorders [80, 81, 89], and five did not provide any information [76, 84, 85, 91, 93]. For the time since stroke event, one study gave no indication [89], two generally less than 3 months [55, 90], two from 3 to 6 months [56, 80], three less than 6 months [57, 59, 87], two 0.5–0.6 months [58, 86], seven 1.0–3.2 months [77, 81, 85, 88, 91, 93, 94], two 4.0–4.9 months [79, 92], one more than 6 months [84], and four 8.3–20.8 months [76, 78, 82, 83]. Six populations were analyzed in China [55, 57, 58, 90,91,92], four in Korea [79, 83, 84, 88], two in Germany [77, 81], two in Italy [56, 87], two in Canada [85, 86], and one each in Pakistan [80], Egypt [76], Iran [78], the Netherlands [93], Finland [94], Russia [59], South Africa [89], and Sweden [82].

Intervention characteristics

In the selection of therapy methods, four studies illustrated general visual skills training [55, 56, 85, 86], four visuomotor training [76, 79, 92, 93], four visual-spatial training [59, 82, 83, 87], three visual feedback training [58, 90, 91], three visual scanning training [80, 81, 94], two eye-movement training [57, 89], and two visual-cognitive training [77, 78]. One study each used vision control dual-task training [84] and eye-tracking visual cognitive training [88] as an intervention approach. All included studies differed significantly in the total period of intervention, number of sessions, and session duration, ranging from 10 days to 10 weeks, eight to 48 sessions, and 20 to 270 min.

Six studies carried out eight to 16 sessions with a total of 300–600 min of visual training [55, 58, 59, 77, 78, 91], 10 conducted 20–25 sessions and 600–4200 min [80,81,82,83,84,85,86,87, 89, 92], and six 28–48 sessions and 840-11520 min [57, 76, 79, 88, 90, 94]. One study did not provide information on the number of sessions or session duration [93] and one did not provide information on session duration [56].

Outcome measures

Of the included studies, nine measured at least one outcome for cognitive function [55, 59, 76, 78, 82,83,84,85, 94], nine only measured activities of daily living [58, 77, 80, 81, 87, 90,91,92,93], and six studies considered both outcomes [56, 57, 79, 86, 88, 89]. In the area of cognitive function, six examined global cognitive function [55,56,57, 59, 76, 89], six executive functions [78, 82,83,84,85, 94], and three considered both outcomes [79, 86, 88]. To measure global cognitive function the MoCA was used in three studies [55, 57, 76], the MMSE in four [56, 79, 88, 89], and two studies considered both tests [59, 86]. The WAIS-DS was used to measure working memory function in five studies [79, 82, 83, 86, 94], the TMT-A for visual processing time in five [79, 83,84,85,86], and the TMT-B for cognitive flexibility in three studies [84,85,86]. The V-CPT was used as a measurement instrument for selective attention in three studies [78, 79, 88] and the ST in two, one of which used the ST to measure selective attention [84] and the other to measure inhibitory control [82]. Table 3 shows all the effects of the visual skills interventions on cognitive and executive function as well as activities of daily living in stroke patients.

Effects of visual skills training on global cognitive function, executive functions and activities of daily living in stroke patients

Figure 2. Effects of visual skills training on global cognitive function (e.g., MMSE) in stroke patients. CI = confidence interval; CON = control group; df = degrees of freedom; INT = intervention group; IV = inverse variance; SE = standard error; Std. = standard.

Fig. 2
figure 2

shows the effects of visual skills training on global cognitive function. In total, the weighted mean SMD resulted in 0.62 (Chi2 = 33.33, df = 8, p < .0001, I2 = 76%), indicating a moderate-sized effect favoring the INT groups

The effects of visual skills training on executive functions in stroke patients are outlined in Fig. 3. The weighted mean SMD amounted to 0.20 (Chi2 = 29.61, df = 17, p = .03, I2 = 43%), which indicates a small-sized effect in favor of the INT groups. Further, a subgroup-analyses revealed small-sized effects for working memory function (SMD = 0.44, Chi2 = 14.71, df = 4, p = .005, I2 = 73%), visual processing speed (SMD = 0.15, Chi2 = 2.91, df = 4, p = .57, I2 = 0%), cognitive flexibility (SMD = 0.27, Chi2 = 3.65, df = 2, p = .16, I2 = 45%), and selective attention (SMD = 0.13, Chi2 = 7.75, df = 4, p = .10, I2 = 48%).

Fig. 3
figure 3

Effects of visual skills training on executive functions (e.g., WSAI-DS) in stroke patients. CI = confidence interval; CON = control group; df = degrees of freedom; INT = intervention group; IV = inverse variance; SE = standard error; Std. = standard

For the effects of visual skills training on activities of daily living in stroke patients, illustrated in Fig. 4, the weighted mean SMD indicated 0.55 (Chi2 = 48.29, df = 14, p < .0001, I2 = 71%), thus revealing a moderate effect in favor of the INT-groups.

Fig. 4
figure 4

Effects of visual skills training on activities of daily living (e.g., MBI) in stroke patients. CI = confidence interval; CON = control group; df = degrees of freedom; INT = intervention group; IV = inverse variance; SE = standard error; Std. = standard

Reporting bias and sensitivity analysis

Funnel Plots are illustrated in Fig. 5A-C. For all measures, the symmetry is limited, indicating a possible publication bias. However, Egger’s test showed no asymmetry for global cognitive function (value = -0.374; p = .709), executive functions (value = 1.058; p = .290), and activities of daily living (value = 0.565; p = .572), respectively. The results therefore give no indication of a publication bias.

Fig. 5
figure 5

Funnel plots for publication bias assessment regarding (A) global cognitive function, (B) executive functions, and (C) activities of daily living. SE: standard error; SMD: standardized mean difference

After finding high heterogeneity (I2 = 76%) regarding global cognitive function, the study by Chen et al. [55] was removed for sensitivity analysis, resulting in a reduced I2 value of 58%. Further removing the study by Rizkalla [86] reduced heterogeneity significantly (I2 = 18%), also revealing a reduction of the weighted mean SMD to 0.32, indicating small-sized effects. An sub-group analysis was conducted to explore the potential mediating effect of time since stroke on global cognitive function scores. The analysis revealed a statistically significant association (p = .027, Table 4) between time since stroke and global cognitive function, suggesting that time since stroke may serve as a moderating factor.

Table 4 Sub-group/moderator analysis of time since stroke and global cognitive function

As a high heterogeneity was found for executive functions in the working memory function subgroup (I2 = 73%), the study by Kim et al. [79] was removed for the sensitivity analysis, which reduced I2 to a value of 39%. The exclusion of Westerberg et al. [82] resulted in a homogeneous outcome (I2 = 0%) and did not alter the effect size (SMD = 0.46). When looking at the study characteristics, the large number of men compared to women was apparent in both studies. A quotient was calculated to determine the proportion of male participants in each study and analyzed as a potential moderator variable. The analysis showed no significant correlation (p = .420) between the average number of male participants and the working memory function (Table 5).

Table 5 Sub-group/moderator analysis of proportion of male participants and working memory function

Description of Image

Source link