Effects of computerized cognitive training on functional activity and connectivity in Huntington's disease: A pilot study

Abstract

Background

Computerized cognitive training (CCT) has been found to improve cognition by altering functional activity and functional connectivity of brain networks in people with and without cognitive impairment. The effects of CCT on functional brain networks in Huntington's disease (HD) have not been comprehensively examined.

Objective

In our pilot trial of CCT, we aimed to explore effects of CCT on functional activity and connectivity of fronto-striatal regions during processing speed and cognitive flexibility tasks, and functional connectivity of resting-state networks in HD.

Methods

Sixteen participants in pre-manifest and early stages of HD were randomised to either a 12-week multi-domain CCT intervention (n = 6) or lifestyle education (n = 10). Participants completed a 1-h magnetic resonance imaging (MRI) scan at baseline and follow-up, which included task-based and resting-state functional MRI. Analyses examined changes in functional activity and connectivity of fronto-striatal regions during processing speed and cognitive flexibility task performance, as well as functional connectivity within default mode and frontoparietal resting-state networks.

Results

While there was evidence of benefits to in-scanner task performance, there were no significant effects on functional activity or functional connectivity of fronto-striatal regions during task performance, or resting-state functional connectivity.

Conclusion

CCT did not generate significant effects on functional activity or connectivity of fronto-striatal networks associated with processing speed or cognitive flexibility, or resting-state networks in HD. A larger study is required to further examine the effects of CCT on functional brain outcomes and potential moderating factors.

Plain language summary

Computerized cognitive training (CCT) or ‘brain training’ has been found to improve cognitive functions by altering brain networks. This includes changing the activation of brain regions, as well as the communication between brain regions. The effects of CCT on brain networks in Huntington's disease (HD), however, have not been comprehensively examined. We conducted a study to determine the effects of CCT on brain networks in HD. 16 participants with pre-symptomatic or early-stage HD were randomly allocated to either a 12-week CCT program, or lifestyle education. Participants completed a 1-h magnetic resonance imaging (MRI) scan at the start and end of the study (after 3 months). This involved scanning the participants while they completed cognitive tasks, and while they were at rest. For each participant, we looked at the activation and communication of brain regions in several key networks. To determine the effects of CCT, we compared whether changes in these networks across the 3-month period differed between the CCT group and the lifestyle education group. We found that although CCT benefited performance on cognitive tasks, there was no impact on brain networks involved in the tasks, or while at rest. Further research is required to confirm the brain changes that underlie the effects of CCT in HD.

Keywords

cognitive training Huntington's disease functional activity functional connectivity cognition

Introduction

Huntington's disease (HD) is a rare, genetically inherited neurodegenerative disease that causes progressive cognitive decline, particularly in processing speed and cognitive flexibility domains in the early stages of the disease.¹ These cognitive changes are associated with the atrophy of fronto-striatal regions and changes in functional connectivity of networks involving these regions.^1–5 Functional connectivity refers to correlations in activity between spatially distinct regions, with greater functional connectivity indicating stronger communication between regions.⁶ Functional connectivity can be examined in the task-free resting-state, or in the task-active state. Resting-state connectivity measures spontaneous brain connectivity, whereas task-related connectivity measures context-dependent connectivity.⁷ Changes in connectivity of task-related functional networks^8–10 as well as resting-state networks such as the default mode network (DMN) and frontoparietal network (FPN) have been found to occur with disease progression in HD.¹¹

Cognitive training has been found to improve cognition in healthy older adults and other neurological conditions by altering functional activity and connectivity of task-related and resting-state networks.^12,13 For example, in older adults and people with multiple sclerosis, computerized cognitive training (CCT) leads to improved performance and increased task-related activation during working memory and response inhibition tasks.^14,15 Increased activation is often interpreted as increased capacity of the network to meet task demands.¹⁶ However, CCT has also been shown to generate improved performance, alongside decreased task-related activation during processing speed and response inhibition tasks in older adults.^17,18 Decreased functional activation is typically interpreted as increased network efficiency, with less neural effort required.¹⁶ These differing effects may be due to population characteristics, with increased activity occurring more often in populations with cognitive impairment, while decreased activity is more common in cognitively healthy populations.¹³ Effects on functional connectivity appear more consistent. Across healthy older adults, people with mild cognitive impairment, and multiple sclerosis, task-related functional connectivity and resting-state functional connectivity within the DMN and FPN typically increases following CCT, which may indicate stronger networks and normalisation of age- or disease-related changes.^13,14,19

Research into the effects of cognitive training on cognition and functional brain outcomes in HD is still in its infancy.²⁰ To our knowledge, only one study has reported effects of CCT on functional brain outcomes. In this study, there was improved cognition and increased functional connectivity between central executive and sensorimotor resting-state networks following CCT in people with early-to-middle stage HD.²¹ However, task-related functional activity/connectivity were not examined, and neuroimaging data were not available for a control group, which limits conclusions regarding the effects of CCT on functional brain outcomes.

We conducted a pilot trial in pre-manifest and early-stage HD, comparing CCT with lifestyle education. We previously reported cognitive, psychosocial,²² and grey matter outcomes²³ from this cohort. Here, we report the effects of CCT on task-related functional brain activity and connectivity, and resting-state connectivity. Here, we focus on effects on functional activity and connectivity of fronto-striatal regions during processing speed and cognitive flexibility tasks, and resting-state connectivity in DMN and FPN, since these cognitive domains and functional networks are affected in HD.^1,5,8,11 We hypothesised increased functional activity and functional connectivity in task- and resting-states in the CCT group from baseline to follow-up, compared to the lifestyle education group.

Methods

The protocol has been published previously.²⁴ Here, we focus on methods relevant to functional neuroimaging outcomes. Further details are provided in Supplementary Material. This study was approved by Monash University Human Research Ethics Committee (approval no. 16420) and conducted in accordance with the World Medical Association Declaration of Helsinki. The study was prospectively registered on the Australian New Zealand Clinical Trials Registry (ACTRN12622000908730).

Participants

We recruited people with pre-manifest or early-stage HD across Australia. People were eligible if they were a) HD gene-positive (CAG repeat length > 35), b) in either pre-manifest or early manifest stage of HD, defined by having a total functional capacity (TFC) score of 7 or above on the Unified Huntington's Disease Rating Scale (UHDRS),²⁵ and c) had access to a computer with internet connection. People were ineligible if they were a) under 18 years of age, b) had a diagnosis of any major neurological or psychiatric conditions other than HD, c) had a history of substance abuse or head injury, d) experiencing severe anxiety or depression symptoms (> 14 on either subscale of Hospital Anxiety and Depression Scale [HADS],²⁶ or e) had an unstable dose of medication for anxiety or depression in the last 6 months. Individuals were excluded for magnetic resonance imaging (MRI) if they were pregnant or had any MRI contraindications. Individuals were initially excluded for left-handedness; however, this criterion was relaxed at later stages of the trial.

Interventions

Intervention: CCT

Intervention group participants completed 2 × 60-min CCT sessions each week, over 12 weeks, using the BrainHQ online platform developed by Posit Science (www.brainhq.com) at home. One session each week was supervised remotely by a researcher through video teleconferencing software. The training schedule involved visual and auditory tasks that target processing speed, attention, working memory, and executive functioning. The program is adaptive, and the difficulty level adjusts to user performance.

Control: lifestyle education

The control group received lifestyle education through three monthly newsletters, which provided information on physical activity, cognitive and social engagement, and diet in relation to brain health. Control group participants received monthly check-in calls and access to CCT for 12 weeks after study participation.

Study procedure

Participants completed assessments with cognitive tests and questionnaires at three timepoints: pre-baseline, baseline, and follow-up. A sub-sample of participants who consented to MRI also completed MRI scanning at baseline and follow-up.

The pre-baseline assessment involved cognitive testing to reduce practice effects, and collection of demographic information. Premorbid intelligence was estimated using National Adult Reading Test (NART).²⁷ Handedness was assessed following relaxation of handedness criterion using Edinburgh Handedness Inventory (EHI).²⁸ Baseline and follow-up assessments involved cognitive tests, questionnaires, and MRI scanning for the MRI sub-sample.

MRI scanning

Imaging data were collected on a 3T Siemens (Erlangen, Germany) Skyra MRI scanner. High-resolution structural T1-weighted magnetization-prepared rapid gradient-echo images were acquired (TR = 2300 ms, TE = 2.07 ms, TI = 900 ms, flip angle 9°, field of view = 256 × 256 mm, 192 slices, bandwidth = 230 Hz/pixel, 1 × 1 × 1 mm voxel size). Functional MRI (fMRI) scans were acquired using echo planar imaging (TR = 2550 ms, TE = 30 ms, flip angle 80°, field of view = 192 × 192 mm, 44 slices, bandwidth = 1778 Hz/pixel, 2.5 × 2.5 × 3 mm voxel size). fMRI scans were acquired during resting state (TA = 5 min), where participants viewed a fixation cross, and while participants completed two tasks in the scanner (TA = 9.43 min × 2 runs of task-switching (TSWT); TA = 9.20 min × 2 runs of modified Symbol Digits Modalities Test [SDMT]).

fMRI tasks

Letter-number TSWT

The cued-trial letter-number TSWT paradigm was used as a measure of cognitive flexibility.²⁹ Participants were presented with a letter-digit pair and switched randomly between completing either a letter task or a number task (Figure 1). In the letter task, participants were required to classify the letter as vowel (A, E, I, U) or consonant (G, K, M, R). For the number task, participants were required to classify the number as odd (3, 5, 7, 9) or even (2, 4, 6, 8). Participants were asked to respond as quickly and as accurately as possible using index fingers and received feedback (tick or cross). Participants completed 72 switch and 72 repeat trials (total 144 trials, 18 min) across two blocks. To facilitate task completion within scanner, all participants completed a practice session on a computer in pre-baseline (136 trials, 17.4 min), baseline (64 trials, 8 min), and follow-up sessions (136 trials, 17.4 min).

Figure 1.

Design of the letter-number TSWT paradigm (a) and modified SDMT (b). In TSWT, participants were cued to complete either the letter task or number task for each letter-number pair. The letter task required classifying the letter as vowel or consonant. The number task required classifying the number as odd or even. In modified SDMT, participants were required to indicate whether a digit-symbol probe matched a digit-symbol pair in a coding table. TSWT: task switching; SDMT: Symbol Digits Modalities Test; ITI: intertrial interval.

Modified SDMT

The modified SDMT was used as a measure of processing speed.³⁰ The task required indicating whether a digit-symbol probe matched a digit-symbol pair in a coding table (Figure 1). Participants were required to respond as quickly as possible using their index fingers and received feedback (tick or cross). Participants completed 72 match and 72 no match trials (total 144 trials, 16.8 min) across two blocks. All participants completed a practice session of the modified SDMT on a computer in pre-baseline (72 trials, 8.4 min), baseline (45 trials, 5.3 min), and follow-up sessions (72 trials, 8.4 min).

Randomisation

Following the baseline session, participants were randomised by an off-site researcher using MinimPy2 software. Minimisation was used to balance the following variables between groups: age, sex, years of education, disease stage (TFC score), severity of anxiety/depression symptoms (HADS score), consent to MRI.

Outcomes

Complete study outcomes are described in our trial protocol.²⁴ Here, we focus on changes in functional brain outcomes in the CCT group, compared to the lifestyle education group. This included task-related functional activity and connectivity of fronto-striatal regions during TSWT and modified SDMT. We also considered changes in resting-state functional connectivity within DMN and FPN.

Analyses of clinical and demographic variables

Statistical analyses were conducted using RStudio (2024.12.0+467). Groups were compared on demographic (age, years of education, premorbid IQ [NART errors], handedness [EHI laterality index]) and clinical variables (CAG length, TFC score, total motor score, disease burden score [DBS],³¹ total HADS score) at baseline using independent samples t-tests. Distributions of sex, handedness (left/right), and disease stage were compared between groups using Chi-squared test.

Analyses of fMRI task performance

Group differences in task performance change were analysed using a 2 (time) × 2 (group) ANOVA interaction. Significant interaction effects were explored with simple effects analyses within groups using repeated measures t-tests. TSWT outcome measures included switch trial accuracy, repeat trial accuracy, accuracy switch cost (repeat accuracy – switch accuracy, with larger values representing greater decrement in accuracy when switching), switch trial reaction time, repeat trial reaction time, reaction time switch cost (switch reaction time – repeat reaction time, with larger values representing slower response when switching). Modified SDMT outcome measures included accuracy and reaction time across trials.

Analyses of neuroimaging data

Pre-processing and analysis of structural MRI data

Structural MRI data were pre-processed using CAT12 toolbox (RRID: SCR_019184) and SPM12 (RRID: SCR_007037). T1 images were segmented, normalised, and modulated using a longitudinal segmentation model optimised for detecting small changes, and smoothed using an 8 mm full width half maximum (FWHM) Gaussian kernel. All T1 images had an image quality rating of <3 (satisfactory) based on CAT12, and were included in analyses. Total grey matter volumes at baseline were compared between groups using an independent samples t-test.

Pre-processing of fMRI data

fMRI data were pre-processed in SPM12 and CONN toolbox (RRID: SCR_009550) using standard pre-processing procedures. Pre-processing included slice timing correction, realignment, co-registration, segmentation, normalisation and smoothing using an 8 mm FWHM Gaussian kernel. Outlier scans were determined using ART outlier detection in CONN, defined as acquisitions with above 0.9 mm framewise displacement or global BOLD signal changes above 5 SD. Quality control measures outputted by the ART outlier detection procedure for each participant during each fMRI paradigm (proportion of valid scans out of total scans, mean global signal change across valid scans, and mean head movement/framewise displacement across valid scans) were compared between groups using independent samples t-tests. fMRI data underwent additional denoising prior to functional connectivity analyses, including band-pass filtering and regression of potential confounding effects (white matter and cerebrospinal fluid signals, motion, outlier scans).

Selection of task-related region of interests (ROIs)

To analyse task-related functional activity and connectivity, a priori ROIs were selected. We chose fronto-striatal ROIs including the caudate, putamen, and DLPFC as these regions are task-relevant^32,33 and implicated in HD pathology.⁵ In our published protocol, we selected ROIs based on task-activated regions reported in meta-analyses of TSWT and modified SDMT.²⁴ However, given the small sample, we reduced the number of ROIs to increase sensitivity for changes in HD-specific regions. For both tasks, 4 mm spherical ROIs within bilateral caudate and putamen were selected from the 300 ROI atlas,³⁴ based on their prominent integration with frontoparietal and ventral attention resting-state networks. Frontoparietal and ventral attention resting-state networks are associated with executive functioning and attentional reorientation, respectively, and are thus engaged by the fMRI tasks.^35,36 5 mm spherical ROIs in the left DLPFC were selected for each task according to coordinates reported in previous meta-analyses of fMRI studies using TSWT and modified SDMT to ensure task-specificity.^37,38 ROI coordinates are provided in Supplementary Material.

Analysis of task-related activity

To analyse task-related functional activity, separate general linear models (GLMs) were defined for each task in SPM12. Details of the GLMs are provided in Supplementary Material. Activation contrasts were estimated for each participant. For TSWT, the activation contrast was switch > rest, and for modified SDMT, the activation contrast was trial > rest. To identify task-activated regions, group-level contrasts were estimated by running one-sample t-tests with all individual-level activation contrasts across timepoints. Statistical significance was set at p < .05, false discovery rate (FDR)-corrected for TSWT, and p < .001, FDR-corrected for SDMT. A more stringent statistical threshold was chosen for SDMT, as very large supra-threshold clusters were obtained at the lower threshold. Clusters were extent-thresholded at k > 25 voxels.

Changes in functional activity between baseline and follow-up were compared between groups using ROI-to-ROI analyses. Eigenvariates for each ROI (defined above) were extracted from individual activation maps at each timepoint and analysed using a 2 (time) × 2 (group) mixed ANOVA interaction. Statistical significance was set at p < .05.

Analysis of task-related connectivity

For each task, changes in functional connectivity between regions associated with task performance were estimated using generalised psychophysiological interactions (gPPI) analyses in CONN. A gPPI model includes seed BOLD signals (physiological factors), boxcar signals characterising task condition convolved with canonical haemodynamic response function (psychological factors), and interaction terms (product of physiological and psychological factors).³⁹ The change in functional connectivity associated with task condition is the regression coefficient of each interaction term (gPPI value).

At the individual level, gPPI models were computed for each pair of seed and target ROIs (defined above). Details of the models are provided in Supplementary Material. At the group level, gPPI values associated with switch condition (relative to rest) in TSWT, and trial condition (relative to rest) in SDMT, were analysed for each pair of ROIs. Changes in task-associated functional connectivity from baseline to follow-up were compared between groups using a 2 (time) × 2 (group) mixed ANOVA interaction. Statistical significance was set at p < .05, FDR-corrected.

Analysis of resting-state functional connectivity

Changes in functional connectivity within DMN and FPN were compared using ROI-to-ROI analyses. ROIs were selected from the Human Connectome Project atlas in CONN.⁴⁰ DMN ROIs included medial prefrontal cortex, precuneus, and bilateral lateral parietal cortex. FPN ROIs included bilateral lateral prefrontal cortex and lateral posterior parietal cortex. ROI coordinates are specified in Supplementary Material.

For each network, at the individual level, functional connectivity values were estimated using a weighted least squares linear model in CONN, where weights represent each session convolved with canonical haemodynamic response function. Connectivity values are Fisher-transformed bivariate correlation coefficients. First-level covariates included realignment parameters and outlier scans. At the group level, changes in ROI-to-ROI connectivity from baseline to follow-up were compared between groups using a 2 (time) × 2 (group) mixed ANOVA interaction. Statistical significance was set at p < .05, FDR-corrected.

Additionally, for each network, connectivity values for each ROI-to-ROI connection were averaged to create an overall network-level measure of connectivity (e.g., overall connectivity of DMN). Changes in overall connectivity from baseline to follow-up were compared between groups using a 2 (time) × 2 (group) mixed ANOVA.

Results

Participants

Sixteen participants consented to MRI and were randomised (six to CCT, ten to lifestyle education). There were no significant group differences on demographic or clinical variables, or total grey matter volumes at baseline (Table 1). One participant in the CCT group was removed from task-related analyses as they reported significant fatigue (unrelated to study) at follow-up.

Table 1.

Baseline demographic and clinical information.

Variable	CCT (n = 6)	Lifestyle education (n = 10)	p-value
Age (years)	45.67 (10.88)	48.60 (14.18)	.45
Sex (% male)	16.67	40.00	.68
Handedness (% right-handed)	100.00	90.00	1.0
EHI Laterality Quotient^a	—	0 (141.42)	—
Education (years)	17.83 (3.86)	16.20 (4.47)	.44
Premorbid functioning (NART errors)	19.50 (6.89)	17.00 (3.94)	.34
CAG repeat length^b	40.00 (1.27)	40.80 (1.93)	.65
Disease stage (% pre-manifest)	83.30	80.00	1.0
UHDRS TFC	12.83 (0.41)	12.20 (1.93)	.16
DBS	205.83 (77.86)	257.70 (112.60)	.30
UHDRS TMS	1.17 (1.60)	5.50 (8.89)	.33
HADS total score	5.67 (4.84)	5.10 (5.02)	.83
Total grey matter volume (mm³)	631.36 (67.62)	646.16 (105.84)	.74

Note: Data are presented as mean (SD) unless stated otherwise. p-value is based on independent samples t-test for continuous variables, and Chi-squared test for categorical variables.

CCT: computerized cognitive training; EHI: Edinburgh Handedness Index; NART: National Adult Reading Test; DBS: disease burden score; UHDRS: Unified Huntington's Disease Rating Scale; TFC: total functional capacity; TMS: total motor score; HADS: Hospital Anxiety and Depression Scale.

Available for n = 2 participants as it was assessed for new participants after relaxing handedness criterion.

Proportion of participants with CAG repeat lengths in 36–39 range was 50% in CCT group and 40% in lifestyle education group.

fMRI quality control measures

Descriptive data and inferential statistics on fMRI quality control measures are provided in Supplementary Material. When comparing groups, proportion of valid scans (usable scans out of total scans) and mean global signal change across valid scans did not significantly differ between groups for any fMRI paradigm (rest, SDMT, TSWT) (p's > .05). Similarly, during SDMT or TSWT paradigms, mean head movement (framewise displacement) across valid scans did not differ between groups. However, during resting-state fMRI, mean head movement across valid scans was significantly higher in the lifestyle education group, compared to the CCT group (p = .02).

In-scanner task performance

Task performance data and statistics are shown in Table 2. For TSWT, there was a significant group×time interaction effect in switch trial accuracy. Simple effects analyses revealed no significant change in switch accuracy in the CCT group (p = .28), and a trend towards decreased switch accuracy in the lifestyle education group (p = .08). Similarly, there was a significant interaction effect in accuracy switch cost, however simple effects analyses showed no significant change in accuracy switch cost within each group (p's > .05). There were no significant interaction effects in switch reaction time, repeat trial accuracy or reaction time, or reaction time switch cost.

Table 2.

Descriptive data and inferential statistics for cognitive outcomes.

Variable	CCT (n = 6)			Lifestyle education (n = 10)			Statistics
Variable	N ^a	T1	T2	N ^b	T1	T2	F	p
TSWT Switch Accuracy (%)	5	95.21 (7.01)	97.78 (2.52)	9	96.40 (3.08)	94.99 (3.53)	5.05	. 04
TSWT Switch Reaction Time (ms)	5	879.89 (185.01)	849.33 (210.43)	9	1002.15 (230.85)	1006.83 (239.45)	0.62	.45
TSWT Repeat Accuracy	5	97.46 (3.05)	96.34 (4.18)	9	97.18 (2.34)	97.34 (3.69)	1.08	.32
TSWT Repeat Reaction Time (ms)	5	851.50 (198.26)	793.62 (152.63)	9	946.11 (238.18)	987.17 (231.64)	4.05	.07
TSWT Switch Cost Accuracy (%)	5	2.25 (4.06)	−1.45 (2.23)	9	0.78 (2.83)	2.35 (2.82)	5.58	. 04
TSWT Switch Cost RT (ms)	5	28.40 (41.91)	55.71 (95.03)	9	56.04 (44.49)	19.65 (43.23)	1.76	.21
mSDMT Accuracy (%)	5	95.14 (3.40)	94.86 (4.57)	10	92.85 (11.88)	93.61 (7.89)	0.22	.65
mSDMT Reaction Time (ms)	5	1311.71 (238.70)	1239.46 (211.25)	10	1583.39 (442.08)	1596.58 (437.23)	5.11	. 04

Note: Data are presented as mean (SD) unless otherwise stated. Bolded p-values are considered statistically significant.

CCT: computerized cognitive training; T1: baseline; T2: follow up; TSWT: task switching; mSDMT: modified Symbol Digits Modalities Test

Cognitive data was removed for 1 participant in the CCT group.

One participant in the lifestyle education group was not able to complete TSWT task. N is representative of number of observations at baseline and follow up.

For modified SDMT, there was a significant interaction effect in trial reaction time. Simple effects analyses revealed significantly improved reaction time in the CCT group (p = .03), and no significant change in the lifestyle education group (p = .60). There was no significant interaction effect in trial accuracy.

Functional activation during in-scanner tasks

Coordinates and test statistics of significant regions of activation during TSWT and modified SDMT are provided in Supplementary Material. For TSWT, group-level analyses revealed significant activation of frontal, parietal, and occipital regions during switch trials, compared to rest (Figure 2). This included bilateral superior frontal gyri, precentral gyri and supplementary motor areas, insula and inferior frontal gyri, left superior parietal lobule, and bilateral inferior occipital gyri.

Figure 2.

(a) brain regions activated during switch condition in TSWT and trial condition in modified SDMT, compared to rest. Significant clusters were false discovery rate (FDR)-corrected at p < .05 for TSWT, and p < .001 for modified SDMT. Clusters were extent-thresholded at k > 25 voxels. (b) Functional activity of selected fronto-striatal ROIs during TSWT. (c) Functional connectivity of selected fronto-striatal ROIs during TSWT. (d) Functional activity of selected fronto-striatal ROIs during modified SDMT. (e) Functional connectivity of selected fronto-striatal ROIs during modified SDMT. CCT: computerized cognitive training; TSWT: task switching; SDMT: Symbol Digits Modalities Test; L: left; R: right; ROIs: regions of interest; DLPFC: dorsolateral prefrontal cortex; T1: baseline; T2: follow up.

For modified SDMT, group-level analyses revealed increased activation of frontal, parietal, and occipital regions during trials, compared to rest (Figure 2). This included bilateral superior frontal gyri, precentral gyri, supplementary motor areas, frontal inferior opercula, middle occipital gyri and superior parietal lobules, insula, and left thalamus.

Effects of CCT on task-related functional activity and connectivity

For TSWT and SDMT, there were no significant interaction effects in functional activity of any fronto-striatal ROIs (p's > .05, uncorrected), or functional connectivity of any ROI-to-ROI connections (p's > .05, FDR-corrected) (Figure 2).

Effects of CCT on resting-state functional connectivity

There were no significant interaction effects in functional connectivity of individual ROI-to-ROI connections (p's > .05, FDR-corrected), or in average within-network connectivity of DMN or FPN (p's > .05, uncorrected) (Figure 3).

Figure 3.

Functional connectivity of selected ROI-to-ROI connections, and average functional connectivity within the DMN (a) and FPN (b). CCT: computerized cognitive training; DMN: default mode network; FPN: frontoparietal network; L: left; R: right; PCC: posterior cingulate cortex; PFC: prefrontal cortex; ROI: region of interest; T1: baseline; T2: follow up.

Inter-individual heterogeneity

Individual-level values of functional activity and functional connectivity (for task-related and resting-state networks) were visualised to further understand the non-significant effects. Qualitative examination indicated heterogeneity in direction of change between baseline and follow-up within each group (Figures 2 and 3). While the small sample prevented statistical analysis of moderating factors, visualisation of datapoints coded by individual factors including age, DBS, and CAG length did not appear to explain inter-individual variability (Supplementary Material).

Discussion

Our pilot study investigated the effects of a 12-week multi-domain CCT intervention on functional activity and connectivity of fronto-striatal networks associated with processing speed and cognitive flexibility, as well as resting-state functional connectivity in pre-manifest and early-stage HD. Compared to the lifestyle education group, participants in the CCT group exhibited improved processing speed and better maintenance of task switching accuracy from baseline to follow-up, based on in-scanner task performance. However, there were no significant effects on functional activity, or task- or resting-state functional connectivity between baseline to follow-up.

The absence of significant effects on functional activity and connectivity in our study contrasts with studies in other populations such as healthy older adults, people with mild cognitive impairment, and multiple sclerosis, which show altered functional activity and connectivity during task and resting states.¹³ This discrepancy may be attributable to differences in neurodegenerative burden. Participants in this study (particularly within MRI cohort) were predominantly in pre-manifest stages of HD, with possibly minimal changes in functional brain outcomes at baseline, thus limiting potential effects of CCT. This argument is consistent with reports of compensatory brain activity in pre-manifest HD individuals closer to disease onset,⁴¹ but not in individuals further from disease onset.⁴² Additionally, significant changes in functional connectivity of DMN or FPN typically only emerge in manifest HD.¹¹ Therefore, CCT may have a greater effect on functional neuroimaging outcomes with greater disease burden.

Technical factors, including run time of resting-state fMRI scan may have also limited sensitivity of results within our small sample. We employed a 5 min resting-state scan using a fixation cross paradigm. Previous research has shown that 5–6 min resting-state runs provide adequate intra-individual reliability of functional connectivity analyses, particularly with fixation paradigms.⁴³ However, a longer acquisition time (to around 13 min) may generate greater reliability and sensitivity of analyses.⁴⁴ This should be considered in further adaptations of our protocol.

Furthermore, it is possible that CCT only affects task-related functional brain activity and connectivity in certain cognitive domains. Studies in other populations have primarily examined effects of CCT on working memory-related functional activity and connectivity.^14,19,45 On the other hand, we examined functional brain outcomes during cognitive flexibility and processing speed tasks. Thus, effects of CCT may not generalise to the cognitive domains that we examined.

Our results also differ from a study of CCT in HD that found increased resting-state connectivity between central executive and sensorimotor networks.²¹ However, differences in analyses limit comparison of results. In that study, there were no neuroimaging data for the non-CCT control group. Also, that study examined whole-brain seed-based resting-state functional connectivity, whereas we examined task-related functional connectivity, and resting-state functional connectivity within DMN and FPN.

We acknowledge that there may be several characteristics of our sample which should be considered when interpreting our results. Firstly, the small sample with unevenly sized groups may have reduced statistical power. Unfortunately, uneven group sizes occurred due to the process of randomisation within a small sample. A minimisation procedure was used for the entire cohort (including those without MRI) to reduce group differences in baseline variables. Although this produced balanced group sizes in the larger cohort, group sizes were uneven in the MRI sub-sample. We also note that while the groups did not significantly differ in DBS and TMS, greater head movement in the lifestyle education group versus CCT group during resting-state (but not task-related) scans may have had subtle influences on those analyses, despite motion correction procedures at the analysis stage.

Considering the discussed limitations of the small and unevenly sized groups, we examined individual-level changes to determine whether non-significant results indicated sub-threshold results of a true effect of CCT (evident as a consistent effect across individuals) or otherwise. Examination of individual-level changes indicated no consistent effect of CCT on functional brain activity or connectivity. It remains possible that a larger study may yield significant results. However, based on our data, we cannot conclude that CCT has a significant effect on task-related functional activity or connectivity of fronto-striatal networks, or resting-state connectivity in pre-manifest and early-stage HD.

It is interesting to consider the present findings in conjunction with our previously published results on grey matter volumes in the same cohort.²³ Using voxel-based morphometry analyses to estimate grey matter volumes, we found significant group by time interaction effects suggesting distinct patterns of change in several brain regions. Post-hoc repeated measures t-tests revealed that the CCT group had increased volumes in the Heschl's gyrus, superior temporal gyrus, middle cingulate, and middle frontal gyrus, and preserved volumes within the putamen, while the lifestyle education group had decreased volumes in the same regions.²³ Additionally, visualisations of individual-level changes revealed general consistency in effects within each group. Detailed methods can be found in the previously published manuscript.²³ The differences in outcomes may be indicative of relatively greater utility of grey matter volumes as a neural biomarker for HD progression. This aligns with the incorporation of striatal volumes as a component of the HD Integrated Staging System (HD-ISS),⁴⁶ whereby decreased striatal volumes, even in the absence of cognitive or functional neuroimaging changes, are indicative of disease progression. Thus, grey matter volumes may be a more reliable biomarker for assessing neural mechanisms of interventions in HD.

The discrepancy between changes in performance, grey matter volumes, and functional brain outcomes, although unexpected, has been documented in other populations, particularly those with multiple imaging timepoints. For example, a study of CCT in older adults found that changes in resting-state functional connectivity (after 9 h of training) disappeared with longer duration of training (36 h), while increases in grey matter volumes persisted and correlated with cognitive improvements.⁴⁷ As such, it is possible that the discrepancy in structural and functional brain outcomes in our study is attributable to our timeframe of training and follow-up.

In summary, CCT had no significant nor consistent effect on functional brain activity or connectivity associated with fronto-striatal networks during cognitive flexibility or processing speed task performance, or resting-state functional connectivity, despite improved cognitive task performance and increased grey matter volumes within the same sample.²³ As this was a pilot study, further studies are required to examine the generalisability of these results in a larger sample. Additionally, further research may consider incorporating blood biomarkers associated with neuropathology and disease progression such as neurofilament light⁴⁸ to further explore effects and mechanisms of CCT in HD.

Supplemental Material

sj-docx-1-hun-10.1177_18796397251399752 - Supplemental material for Effects of computerized cognitive training on functional activity and connectivity in Huntington's disease: A pilot study

Supplemental material, sj-docx-1-hun-10.1177_18796397251399752 for Effects of computerized cognitive training on functional activity and connectivity in Huntington's disease: A pilot study by Katharine Huynh, Nellie Georgiou-Karistianis, Amit Lampit, M Navyaan Siddiqui, Katharina Voigt, Julie C Stout and Sharna D Jamadar in Journal of Huntington's Disease

Footnotes

Acknowledgements

We thank all participants and their families for their time and contribution to the study. We also thank the staff and clinicians at Calvary Bethlehem Hospital and Neuropsychiatry Centre, Royal Melbourne Hospital, for their assistance in introducing patients to the study.

ORCID iDs

Katharine Huynh

Nellie Georgiou-Karistianis

M Navyaan Siddiqui

Ethical considerations

This study was approved by the Monash University Human Research Ethics Committee (approval no. 16420). This research was conducted ethically in accordance with the World Medical Association Declaration of Helsinki.

Consent to participate

All participants provided written informed consent to participate prior to enrolment in the study.

Consent for publication

Not applicable.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by Monash University [PAG19-0576] and Huntington's Victoria [2022 Peter Walsh Scholarship]. K.H. and M.N.S are supported by Australian Government Research Training Program (RTP) Scholarship. S.D.J. is supported by an Australian National Health and Medical Research Council (NHMRC) Fellowship. A.L. is supported by the University of Melbourne. N.G-K and J.C.S. received no financial support for the research, authorship, and/or publication of this article.

Declaration of conflicting interests

The authors declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: Professor Julie Stout is an Editorial Board Member of this journal but was not involved in the peer-review process of this article nor had access to any information regarding its peer-review.

Data availability

Individual data has not been shared due to the higher risk of participant identification associated with the low incidence of HD.

Supplemental material

Supplemental material for this article is available online.

References

Tabrizi

Scahill

Owen

, et al. Predictors of phenotypic progression and disease onset in premanifest and early-stage Huntington’s disease in the TRACK-HD study: analysis of 36-month observational data. Lancet Neurol 2013; 12: 637–649.

Dominguez D

Poudel

Stout

, et al. Longitudinal changes in the fronto-striatal network are associated with executive dysfunction and behavioral dysregulation in Huntington’s disease: 30 months IMAGE-HD data. Cortex 2017; 92: 139–149.

Harrington

Rubinov

Durgerian

, et al. Network topology and functional connectivity disturbances precede the onset of Huntington’s disease. Brain 2015; 138: 2332–2346.

Polosecki

Castro

Rish

, et al. Resting-state connectivity stratifies premanifest Huntington’s disease by longitudinal cognitive decline rate. Sci Rep 2020; 10: 1252.

Langley

Gregory

Osborne-Crowley

, et al. Fronto-striatal circuits for cognitive flexibility in far from onset Huntington’s disease: evidence from the young adult study. J Neurol Neurosurg Psychiatry 2021; 92: 143.

Friston

. Functional and effective connectivity: a review. Brain Connect 2011; 1: 13–36.

Masharipov

Knyazeva

Korotkov

, et al. Comparison of whole-brain task-modulated functional connectivity methods for fMRI task connectomics. Commun Biol 2024; 7: 1–21.

Wolf

Sambataro

Vasic

, et al. Aberrant connectivity of lateral prefrontal networks in presymptomatic Huntington’s disease. Exp Neurol 2008; 213: 137–144.

Wolf

Sambataro

Vasic

, et al. Altered frontostriatal coupling in pre-manifest Huntington’s disease: effects of increasing cognitive load. Eur J Neurol 2008; 15: 1180–1190.

10.

Georgiou-Karistianis

Poudel

Domínguez D

, et al. Functional and connectivity changes during working memory in Huntington’s disease: 18 month longitudinal data from the IMAGE-HD study. Brain Cogn 2013; 83: 80–91.

11.

Pini

Jacquemot

Cagnin

, et al. Aberrant brain network connectivity in presymptomatic and manifest Huntington’s disease: a systematic review. Hum Brain Mapp 2020; 41: 256–269.

12.

Duda

Sweet

. Functional brain changes associated with cognitive training in healthy older adults: a preliminary ALE meta-analysis. Brain Imaging Behav 2020; 14: 1247–1262.

13.

van Balkom

van den Heuvel

Berendse

, et al. The effects of cognitive training on brain network activity and connectivity in aging and neurodegenerative diseases: a systematic review. Neuropsychol Rev 2020; 30: 267–286.

14.

Iordan

Cooke

Moored

, et al. Neural correlates of working memory training: evidence for plasticity in older adults. Neuroimage 2020; 217: 116887.

15.

Filippi

Riccitelli

Mattioli

, et al. Multiple sclerosis: effects of cognitive rehabilitation on structural and functional MR imaging measures–an explorative study. Radiology 2012; 262: 932–940.

16.

Zhang

Sauce

. Efficiency and capacity mechanisms can coexist in cognitive training. Nat Rev Psychol 2023; 2: 127–127.

17.

Kühn

Lorenz

Weichenberger

, et al. Taking control! structural and behavioural plasticity in response to game-based inhibition training in older adults. Neuroimage 2017; 156: 199–206.

18.

Ross

Webb

Whitaker

, et al. The effects of useful field of view training on brain activity and connectivity. J Gerontol B Psychol Sci Soc Sci 2019; 74: 1152–1162.

19.

Gozdas

Avelar-Pereira

Fingerhut

, et al. Long-term cognitive training enhances fluid cognition and brain connectivity in individuals with MCI. Transl Psychiatry 2024; 14: 447.

20.

Huynh

Nategh

Jamadar

, et al. Cognition-oriented treatments and physical exercise on cognitive function in Huntington’s disease: a systematic review. J Neurol 2022; 270: 1857–1879.

21.

Moreu-Valls

Puig-Davi

Martinez-Horta

, et al. A randomized clinical trial to evaluate the efficacy of cognitive rehabilitation and music therapy in mild cognitive impairment in Huntington’s disease. J Neurol 2025; 272: 202.

22.

Huynh

Jamadar

Lampit

, et al. Feasibility and effects of cognitive training on cognition and psychosocial function in Huntington’s disease: a randomised pilot trial. J Neurol 2025; 272: 159.

23.

Huynh

Georgiou-Karistianis

Lampit

, et al. Computerized cognitive training increases gray matter volumes in Huntington’s disease: a pilot study. Mov Disord 2024; 39: 2119–2121.

24.

Huynh

Jamadar

Stout

, et al. Effects and mechanisms of computerized cognitive training in Huntington’s disease: protocol for a pilot study. Neurodegener Dis Manag 2024; 14: 203–216.

25.

Huntington Study Group. Unified Huntington’s disease rating scale: reliability and consistency. Mov Disord 1996; 11: 136–142.

26.

Zigmond

Snaith

. The hospital anxiety and depression scale. Acta Psychiatr Scand 1983; 67: 361–370.

27.

Nelson

. National Adult Reading Test (NART). 1982.

28.

Oldfield

. The assessment and analysis of handedness: the Edinburgh inventory. Neuropsychologia 1971; 9: 97–113.

29.

Meiran

Chorev

Sapir

. Component processes in task switching. Cogn Psychol 2000; 41: 211–253.

30.

Rypma

Berger

Prabhakaran

, et al. Neural correlates of cognitive efficiency. Neuroimage 2006; 33: 969–979.

31.

Penney Jr

Vonsattel

Macdonald

, et al. CAG Repeat number governs the development rate of pathology in Huntington’s disease. Ann Neurol 1997; 41: 689–692.

32.

Stiers

Goulas

. Task-specific subnetworks extend from prefrontal cortex to striatum. Cortex 2022; 156: 106–125.

33.

Curtin

Ayaz

Tang

, et al. Enhancing neural efficiency of cognitive processing speed via training and neurostimulation: an fNIRS and TMS study. Neuroimage 2019; 198: 73–82.

34.

Seitzman

Gratton

Marek

, et al. A set of functionally-defined brain regions with improved representation of the subcortex and cerebellum. Neuroimage 2020; 206: 116290.

35.

Uddin

Yeo

BTT

Spreng

. Towards a universal taxonomy of macro-scale functional human brain networks. Brain Topogr 2019; 32: 926–942.

36.

Corbetta

Patel

Shulman

. The reorienting system of the human brain: from environment to theory of mind. Neuron 2008; 58: 306–324.

37.

Jamadar

Thienel

Karayanidis

. Task switching processes. Brain Mapp Encycl Ref 2015; 3: 327–335.

38.

Silva

PHR

Spedo

Barreira

, et al. Symbol digit modalities test adaptation for magnetic resonance imaging environment: a systematic review and meta-analysis. Mult Scler Relat Disord 2018; 20: 136–143.

39.

McLaren

Ries

, et al. A generalized form of context-dependent psychophysiological interactions (gPPI): a comparison to standard approaches. Neuroimage 2012; 61: 1277–1286.

40.

Smith

Beckmann

Andersson

, et al. Resting-state fMRI in the human connectome project. Neuroimage 2013; 80: 144–168.

41.

Klöppel

Gregory

Scheller

, et al. Compensation in preclinical Huntington’s disease: evidence from the track-on HD study. EBioMedicine 2015; 2: 1420–1429.

42.

Soloveva

Jamadar

Velakoulis

, et al. Brain compensation during visuospatial working memory in premanifest Huntington’s disease. Neuropsychologia 2020; 136: 107262.

43.

Van Dijk

KRA

Hedden

Venkataraman

, et al. Intrinsic functional connectivity as a tool for human connectomics: theory, properties, and optimization. J Neurophysiol 2010; 103: 297–321.

44.

Birn

Molloy

Patriat

, et al. The effect of scan length on the reliability of resting-state fMRI connectivity estimates. Neuroimage 2013; 83: 550–558.

45.

Giehl

Ophey

Hammes

, et al. Working memory training increases neural efficiency in Parkinson’s disease: A randomized controlled trial. Brain Commun 2020; 2: fcaa115.

46.

Tabrizi

Schobel

Gantman

, et al. A biological classification of Huntington’s disease: the integrated staging system. Lancet Neurol 2022; 21: 632–644.

47.

Lampit

Hallock

Suo

, et al. Cognitive training-induced short-term functional and long-term structural plastic change is related to gains in global cognition in healthy older adults: a pilot study. Front Aging Neurosci 2015; 7: 14–14.

48.

Parkin

Thomas

Corey-Bloom

. Mapping neurodegeneration across the Huntington’s disease spectrum: a five-year longitudinal analysis of plasma neurofilament light. eBioMedicine 2024; 104: 105173.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

2.09 MB