Longitudinal data on cortical thickness before and after working memory training.

The data and supplementary information provided in this article relate to our research article "Task complexity and location specific changes of cortical thickness in executive and salience networks after working memory training" (Metzler-Baddeley et al., 2016) [1]. We provide cortical thickness and subcortical volume data derived from parieto-frontal cortical regions and the basal ganglia with the FreeSurfer longitudinal analyses stream (http://surfer.nmr.mgh.harvard.edu [2]) before and after Cogmed working memory training (Cogmed and Cogmed Working Memory Training, 2012) [3]. This article also provides supplementary information to the research article, i.e., within-group comparisons between baseline and outcome cortical thickness and subcortical volume measures, between-group tests of performance changes in cognitive benchmark tests (www.cambridgebrainsciences.com [4]), correlation analyses between performance changes in benchmark tests and training-related structural changes, correlation analyses between the time spent training and structural changes, a scatterplot of the relationship between cortical thickness measures derived from the occipital lobe as control region and the chronological order of the MRI sessions to assess potential scanner drift effects and a post-hoc vertex-wise whole brain analysis with FreeSurfer Qdec (https://surfer.nmr.mgh.harvard.edu/fswiki/Qdec [5]).

Specifications Table

Value of the data

Transparency and comparability of research results.

Calculation of effect sizes for future apriori sample size and power calculations.

Information about potential confounding factors that may affect the interpretation of similar studies.

Data

We provide data on cortical thickness and subcortical volume in regions of interest of cognitive control networks [1]. These structural data were acquired on a 3 Tesla GE Magnetic Resonance Imaging (MRI) system [1] and were derived with the FreeSurfer longitudinal analyses stream [2].

Experimental design, materials and methods

48 healthy participants (19–40 years of age) were pseudo-randomly (with the provision to match groups for age and sex) allocated to an adaptive training or an active control group [1]. Both groups underwent structural MRI scanning on a 3 Tesla GE system at the Cardiff University Brain Research Imaging Center (CUBRIC) as well as cognitive assessment [4] before and after two months of working memory training (40 sessions in total) [3]. The training group performed the working memory tasks in an adaptive way, i.e., training demands increased with performance levels whilst the control group performed the same tasks but in a non-adaptive way, i.e., task difficulty was held at a low level and never exceeded an item span of 3. Participants could train from home and their progress and compliance was monitored throughout the training time. Eight participants dropped out so the final sample size for both groups was n=20 each (n=40 in total). Training-related changes in working memory span and executive functioning as well as in cortical thickness and subcortical volume in regions of interest of cognitive control networks (executive, salience, basal ganglia networks) were assessed. Cortical thickness and subcortical volume indices were derived with the FreeSurfer longitudinal analyses stream [2]. Training-specific effects were investigating with group by time interaction effects in the structural and cognitive outcome measures [6], [7] (Tables 1 and 2). Brain-function relationships and potential artefacts in the MRI data due to scanner drift effects were studied with correlational analyses (Fig. 1 and Tables 3 and 4).

Table 1

The results of the post-hoc paired t-tests [t(19)-value and p-value in brackets] comparing average cortical thickness indices before and after the two months training for each group (adaptive training and control group) separately.

	Training (n=20)	Control (n=20)
	t(19)-statistic (p-value)
Right pars triangularis	0.89 (0.38)	4.1 (0.001)
Right pars opercularis	1.44 (1.67)	1.5 (0.14)
Right caudal middle frontal	2.26 (0.03)	1.1 (0.29)
Left pallidum	2.53 (0.026)	0.36 (0.72)
Right insula	2.28 (0.03)	0.29 (0.77)
Left anterior cingulate	1.9 (0.07)	0.95 (0.35)

Table 2

Summary of non-significant results of independent t-tests of absolute changes (difference scores between post and pretraining performance scores) in performance in cognitive benchmark tests.

	t(38)-value	p-value
Double trouble	0.48	0.64
Grammatical reasoning	−1.13	0.26
Tree task	−0.65	0.52
Odd one out	0.71	0.48
Self-ordered search	0.15	0.88
Automated symmetry span	1.16	0.25

Fig. 1

Plots the relationship between average cortical thickness measures in the right occipital lobe as a control region that was not expected to change with the intervention and the chronological order of the acquired MRI scans. There was no evidence of a drift in scanner acquisition across the MRI sessions.

Table 3

Spearman׳s rho correlation coefficient ρ (p-values) between performance changes in the backwards digit span and spatial span tasks and changes in cortical thickness in the right caudal middle frontal gyrus, the right pars triangularis and the right insula and changes in subcortical volume in the left pallidum for the training and the control group (n=20).

	Backwards digit span	Spatial span
Training group
Right caudal middle frontal	0.03 (0.9)	0.02 (0.92)
Right pars triangularis	−0.11 (0.64)	−0.01 (0.95)
Right insula	−0.63 (0.003)	0.27 (0.24)
Left pallidum	−0.32 (0.17)	0.37 (0.11)
Control group
Right caudal middle frontal	0.32 (0.16)	−0.08 (0.74)
Right pars triangularis	−0.09 (0.69)	0.32 (0.16)
Right insula	−0.18 (0.43)	−0.11 (0.66)
Left pallidum	−0.11 (0.64)	−0.18(0.45)

Table 4

The Pearson correlation coefficients r (p-value) between the average time spent on training and changes in cortical thickness/subcortical volume across all regions of interest. There were no significant correlations at Bonferroni corrected level of significance (p<0.0015).

Pearson correlation coefficient r (p-value)	Average active time per training session
Change in cortical thickness in ROIs on left hemisphere
Caudal anterior cingulate	0.041 (0.8)
Caudal middle frontal	−0.028 (0.86)
Inferiorparietal	0.234 (0.15)
Parsopercularis	0.202 (0.21)
Parsorbitalis	0.232 (0.15)
Parstriangularis	0.122 (0.45)
Rostral anterior cingulate	−0.047 (0.77)
Rostral middle frontal	0.210 (0.19)
Superior frontal	−0.036 (0.83)
Superior parietal	0.154 (0.34)
Supramarginal	0.115 (0.48)
Insula	0.092 (0.57)
Change in cortical thickness ROIs on right hemisphere
Caudal anterior cingulate	−0.0289 (0.07)
Caudal middle frontal	0.248 (0.12)
Inferiorparietal	0.321 (0.04)
Parsopercularis	0.099 (0.54)
Parsorbitalis	0.205 (0.20)
Parstriangularis	0.208 (0.19)
Rostral anterior cingulate	0.042 (0.79)
Rostral middle frontal	0.109 (0.50)
Superior frontal	0.177 (0.27)
Superior parietal	0.198 (0.22)
Supramarginal	0.095 (0.55)
Insula	−0.194 (0.23)
Change of subcortical volume in ROIs on left hemisphere
Thalamus	0.114 (0.49)
Caudate	0.108 (0.51)
Putamen	0.207 (0.19)
Pallidum	−0.194 (0.23)
Changes of subcortical volume in ROIs on right hemisphere
Thalamus	0.398 (0.01)
Caudate	0.239 (0.14)
Putamen	0.201 (0.21)
Pallidum	−0.114 (0.48)

Vertex-wise whole brain analysis of group and time effects on cortical thickness in FreeSurfer Qdec [5]

We conducted a post-hoc whole brain vertex-wise analyses in Qdec to test for the effects of group (adaptive training versus active control group) and time of assessment effects (baseline versus outcome) on cortical thickness measures. Cortical thickness indices were derived from the T1-weighted anatomical images smoothed with a kernel of 10. Multiple comparisons were controlled with a False Discovery Rate (FDR) of 5%. There were regions on both hemispheres with significant clusters of group effects across time (see Fig. 2) but no region demonstrated main effects of time or interaction effects between time and group, in either the right or the left hemisphere.

Fig. 2

Displays lateral views on the right and left cortical surface respectively with clusters of regions for which a group effect across time was detected.

Subject area	Psychology
More specific subject area	Cognitive Neuroscience, Brain plasticity, Cognitive training
Type of data	Tables, figures, images of results of qdec analysis
How data was acquired	Magnetic resonance imaging and cognitive assessment. Cortical thickness and subcortical volumes derived with FreeSurfer.
Data format	Analyzed
Experimental factors	Longitudinal randomized controlled intervention study
Experimental features	Longitudinal randomized controlled intervention study comparing the effects of adaptive working memory training relative to non-adaptive control activities (n=20 healthy adults per group) on cognition and MRI derived cortical thickness indices in cognitive control networks.
Data source location	Cardiff University Brain Research Imaging Centre, Cardiff, UK
Data accessibility	Data is provided in this article