Skip to main content
NCBI home page
Aging Brain logoLink to Aging Brain
. 2024 Feb 13;5:100109. doi: 10.1016/j.nbas.2024.100109

Neural correlates of home-based intervention effects on value-based sequential decision-making in healthy older adults

Kathleen Kang a,⁎, Daria Antonenko b, Franka Glöckner a, Agnes Flöel b,c, Shu-Chen Li a,d
PMCID: PMC10876581  PMID: 38380149

Highlights

  • •

    Better performing older adults improved after 5 out of 18 home-practice sessions.

  • •

    Better performing older adults showed delay-dependent DLPFC recruitment at post-intervention.

  • •

    Correlation between DLPFC activity and intervention gains differ between groups.

Keywords: Sequential decision-making, Cognitive intervention, fNIRS, Value-based learning, Aging, Dorsolateral prefrontal cortex

Abstract

Older adults demonstrate difficulties in sequential decision-making, which is partly attributed to under-recruitment of prefrontal networks. It is, therefore, important to understand the mechanisms that may improve this ability. This study investigated the effectiveness of an 18-sessions, home-based cognitive intervention and the neural mechanisms that underpin individual differences in intervention effects. Participants were required to learn sequential choices in a 3-stage Markov decision-making task that would yield the most rewards. Participants were assigned to better or worse responders group based on their performance at the last intervention session (T18). Better responders improved significantly starting from the fifth intervention session while worse responders did not improve across all training sessions. At post-intervention, only better responders showed condition-dependent modulation of the dorsolateral prefrontal cortex (DLPFC) as measured by fNIRS, with higher DLPFC activity in the delayed condition. Despite large individual differences, our data showed that value-based sequential-decision-making and its corresponding neural mechanisms can be remediated via home-based cognitive intervention in some older adults; moreover, individual differences in recruiting prefrontal activities after the intervention are associated with variations in intervention outcomes. Intervention-related gains were also maintained at three months after post-intervention. However, future studies should investigate the potential of combining other intervention methods such as non-invasive brain stimulation with cognitive intervention for older adults who do not respond to the intervention, thus emphasizing the importance of developing individualized intervention programs for older adults.

Introduction

Many daily actions or decisions unfold across a series of steps, which demand the integration of information about action-outcome relations across sequential states in order to act adaptively to achieve short- and long-term goals. This form of sequential action-outcome relations that is often encountered in daily life reflects the Markov decision process, in which an action at a particular state will not only affect the outcome at that particular state, but also influences the transition to the next state. In a task that incorporates such Markov decision process, activity in the lateral orbitofrontal cortex was associated with immediate rewards, while activity in the dorsolateral prefrontal cortex (DLPFC) was associated with delayed rewards that entailed a lag between current actions and outcomes from subsequent decision states [28]. We previously discovered profound age-related impairments in decision performance (accuracy and decision time) in the delayed condition of such sequential decisions, which was accompanied by a reduced BOLD activity in the lateral and medial PFC regions in older adults [11]. This finding implies that under-recruitment of the DLPFC contributes to older adults’ difficulties in learning sequential transition structures to predict long-term outcomes.

Nevertheless, age-related impairments in decision-making could be remediated by cognitive intervention. To date, research on cognitive interventions has generally shown performance gains on trained tasks (see [29] for review). A recent study [1] showed that 9 sessions (across 3 weeks) of intervention resulted in performance gains in working memory and sequential decision-making tasks. However, the potential effects of intervention on task-related neural activity and the associated inter-individual differences were not investigated. Individual differences in intervention-induced performance improvements have attracted considerable interest in the past decade [4], [13] but it remains to be determined as to who benefits and who does not benefit from intervention. Since current research on individual differences contributing to intervention success yielded inconsistent findings [23], brain correlates in cognitive intervention studies should be included to elucidate the factors underlying inter-individual differences in intervention effects. Apart from utilizing brain measures to elucidate individual differences in intervention effects, we also explored whether these differences may also inform between-person variations in maintaining intervention effects. Thus far, research into this matter is even scarcer. In a study in patients with multiple sclerosis, changes in resting state functional connectivity from pre- to post-intervention in the default mode network and anterior cingulate cortex were associated with attention and executive function measured 6 months after the intervention [19]. Functional connectivity predicted the persistence of neuropsychological or behavioral improvements at follow-up testing in psychiatric populations [26], [27].

Taken together, this study aimed to investigate the neural correlates of inter-individual differences in intervention-related performance changes during sequential decision-making in older adults by measuring hemodynamic responses in DLPFC using functional near infrared spectroscopy (fNIRS). We examined differences in task-related brain activity in the DLPFC between those who benefitted and those who did not benefit from a 6-week (18 session) home-based intervention. We adopted a home-based training approach as it reduces the burden on participants to complete several lab visits over a few weeks, thus reducing any sampling bias due to mobility. Given that DLPFC is involved in sequential decision-making [28], [11] and performance level in working memory tasks is known to modulate brain activity [16], we hypothesized that post-intervention DLPFC activity would differ between better responders and worse responders. Furthermore, we explore the relationship between post-intervention DLPFC activity and intervention-induced performance gains. Another exploratory part of our study also aims to investigate whether individual differences in intervention-related behavioral effects and brain activity would be associated with maintenance effects of practice-related gains at a later follow-up session in healthy older adults. As an exploratory measure of whether intervention-related effects would persist at a later time point as compared to baseline, maintenance effects were measured at a follow-up session which was held 3 months after the post-intervention session [8], [32]. We expected that post-intervention DLPFC activity will be correlated with maintenance effects, only in participants who benefited from the intervention.

Methods

The effective sample analysed included 40 healthy older adults (19 females, 21 males, mean age = 67.93 years, SD = 1.47 years, range = 65 – 70 years old) who took part in a home-based, web-based sequential decision-making intervention adapted from a previous cognitive intervention study of working memory and decision-making (e.g., [1], [2]). Participants completed baseline assessment, 6 weeks of home-based practice with a sequential decision-making task (cf. [11]) three times per week on their personal laptops, a post-intervention session where brain hemodynamic responses during task performance were assessed with functional near infrared spectroscopy (fNIRS), and a 3-months follow-up session (see Fig. S1 and text in Supplementary Information for details of participant recruitment, sample characteristics, and intervention protocol). To examine inter-individual differences in intervention effects, we conducted a median split of the sample based on performance on the last intervention session (T18). As such, the participants were either in the ‘Better responders’ group or ‘Worse responders’ group. Median split was used instead of regression because the data was not normally distributed which violates the assumption of regression analysis. Albeit suggestions that median split analyses could result in a loss of power and an increase in Type 1 errors, median split procedures are potentially useful in situations when there is a potential moderator of other relationships (such as brain-behaviour relationships) [15]. Furthermore, median splits do not typically inflate Type 1 errors if the independent variables are uncorrelated as is the case in our study [12]. The study was approved by the Ethics Committee of Technische Universität Dresden (Ref: SR-EK-338072020).

The 3-stage Markov sequential decision-making task has been described in detail elsewhere (e.g., [11]. In short, participants learned choice-outcome associations across three states to maximize rewards. This task typically consists of two reward conditions, immediate and delayed. In the immediate reward condition, the optimal sequence of action would result in consistent small amount of gains across all states whereas in the delayed reward condition, the optimal sequence of action would result in losses at the first two states before a large reward at the third state. The task is particularly demanding under the delayed condition, in which associations between the choice at the current stage and outcome at two stages later must be learned. At baseline, post-intervention and follow-up assessments, decision performance was assessed under immediate and delayed conditions. During the home-based sessions, only the cognitively demanding condition (i.e., delayed condition) was practiced (see text and Fig. S2A in Supplementary Information for task details and procedure).

Fig. 2.

Fig. 2

Open in a new tab

A) HbO responses in the ROI as a function of group and condition. Interaction between reward condition and group, *** denotes p < 0.001 while + denotes marginally significant difference with p < 0.1. Error bars represent ± 1 standard error of mean. Topographical plots showing condition-related effects in B) worse responders and C) better responders; as well as group effects on D) immediate reward condition and E) delayed reward condition. * represents thresholded p-values but none of the individual channels survived correction for multiple comparisons. Colour scale denotes t-values.

Brain correlates of intervention-related effects were assessed at the post-intervention session using a continuous-wave fNIRS system NIRSport (NIRx medical Technologies, LLC, USA). Similar to a previous fNIRS study of sequential decision-making [34], the montage used in this study consisted of 8 sources and 8 detectors, amounting to 18 channels over the left and right DLPFC regions (see text and Fig. S2B in Supplementary Information for details of data acquisition with fNIRS and montage).

Data analysis

Behavioral data with respect to the proportion of optimal choice performance was computed and analyzed using MATLAB R2018b (Mathworks Inc, Natick, MA, USA) and RStudio 4.1.1. Performance across 18 sessions between High and Worse responders was analyzed using ANOVA and single-trial generalized linear mixed-effects models (1 = optimal choice, 0 = suboptimal choice), with Group as between-subject factor and Session as within-subject factor. FNIRS data were preprocessed using nirsLAB (NIRx Medical Technologies, Glenhead, [31]. We focused the analyses on HbO, because it is more reliable than HbR (e.g., [17]. HbO data (i.e., mean GLM Beta weights) was analysed using linear mixed-effects with ‘Condition’ (Immediate Reward, Delayed Reward) as within-subjects factor and ‘Group’ (Better responders, Worse responders) as between-subjects factor, as well as subjects and fNIRS channels nested into subjects as random effects model to account for between-subject variability in hemodynamic concentration changes across channels (e.g., [34]) using the nlme and lme4 package in R. Where the residuals were not normally distributed, we ran permutation ANOVAs using the permanova [14] package (number of permutations = 5000). Where permutation ANOVA was conducted, we report the permutated p-values (pperm) as well. In the case of significant interactions, post-hoc tests were conducted and p-values adjusted for multiple comparisons using Holm correction were reported (padj). Furthermore, brain-behavioural correlations (i.e., Pearson’s correlation) between performance gains, performance maintenance and delay-dependent recruitment of HbO responses in DLPFC were conducted separately for the High and Low Performer Groups (see text in Supplementary Information for details about statistical analyses).

Results

Behavioural results

At baseline, both performer groups did not differ in sequential decision-making performance (main effect of ‘Group’ [χ2 (1) = 0.076, p = 0.78], ‘Group’ × ‘Condition’ interaction [χ2 (1) = 1.49, p = 0.22]). Furthermore, the groups did not differ in any of the cognitive covariates (see Supplementary Table 1).

Across all home-based intervention sessions which focused on the delayed reward condition only, there were significant main effects of ‘Session’, χ2 (17) = 93.34, p < 0.001 and ‘Group’, χ2 (1) = 13.55, p = 0.0002, as well as a ‘Group’ × ‘Session’ interaction, χ2 (17) = 1125.29, p < 0.001. Better responders performed significantly better starting from the fifth intervention session as compared to the first intervention session (all psadj < 0.05, ORs > 3.9). Better responders performed significantly more accurately than worse responders starting from the fourth intervention session onwards [all psadj < 0.01; ORs > 3, see Fig. 1]. Worse responders did not significantly improve even at the final session, as compared to their performance at the first session. At post-intervention, there was a significant main effect of ‘Condition’, χ2 (1) = 11.68, p = 0.0006 in which performance was significantly more accurate in the ‘Immediate Reward’ condition as compared to the ‘Delayed Reward’ condition. The crucial ‘Group’ × ‘Condition’ interaction was also significant, χ2 (1) = 8.94, p = 0.0028. Better responders performed significantly better than worse responders in the delayed reward condition, padj = 0.0065, OR = 3.43. Furthermore, performance on the delayed condition was significantly worse than the immediate condition only in the worse responders group, padj < 0.0001, OR = 0.23. For maintenance effects, at follow-up, there was a significant main effect of ‘Group’, χ2 (1) = 3.97, p = 0.046, revealing that better responders still performed significantly better than worse responders after 3 months post-intervention. There was a significant main effect of ‘Session’, χ2 (1) = 31.98, p < 0.001, indicating that performance at follow-up was significantly better than at baseline. There was, however, no significant ‘Group’ × ‘Session’ interaction, χ2 (1) = 1.24, p = 0.27.

Fig. 1.

Fig. 1

Open in a new tab

Performance accuracy in higher and lower performers in the delayed reward condition at baseline, across the 18 intervention sessions, post-intervention, and follow-up sessions. n.s. represents no statistically significant difference. Error bars represent ± 1 standard error of mean.

fNIRS results at post-intervention

As for HbO activity at post-intervention, the delayed reward condition recruited a higher level of DLPFC activity than the immediate condition, F(1, 650) = 9.11, p = 0.0026, pperm = 0.004. Better responders also recruited DLPFC regions more than worse responders, F(1, 35) = 2.65, p = 0.11, pperm = 0.03. Crucially, there was also a significant ‘Group’ × ‘Condition’ interaction, F(1, 650) = 4.37, p = 0.037, pperm = 0.035 (see Fig. 2) indicating that the effect of condition-dependent modulation, with an overall higher DLPFC recruitment for delayed as compared to immediate reward condition, was only observed in better responders, (padj = 0.0007, d = 0.27) but not for worse responders (padj = 0.61, d = 0.035). When ‘Hemisphere’ was included as a factor, there was no significant main effect or interaction with the factor ‘Hemisphere’, Fs < 0.5, ps > 0.5.

Brain-behavior correlations

Regarding brain-behavior relations, correlational analyses between DLPFC activity assessed at post-intervention and performance gains, adjusting for individual differences in initial performance (i.e., a ratio score computed as T18-T1/T1) revealed a trend towards a significant correlation for better responders, r(18) = 0.41, p = 0.069 but not for worse responders, r(14) = -0.35, p = 0.18 (rho(14) = -0.41, p = 0.11 due to non-normality for worse responders; see Fig. 3). The brain-behavior correlations differed significantly between the two groups, z = 2.17, p = 0.03.

Fig. 3.

Fig. 3

Open in a new tab

Correlation between mean GLM beta weights at DLPFC and learning rate for A) better responders and B) worse responders. Shaded area represents the 95% confidence interval.

As for whether intervention-related performance gains were associated with maintenance effects in relation to baseline performance, no significant correlation was found for both worse responders, rho = 0.36, p = 0.13 and better responders, rho = -0.02, p = 0.92. Regarding whether DLPFC activity at post-intervention was associated with maintenance effects (i.e., a ratio score computed as (follow-up – baseline)/baseline), no significant correlations were found for both groups (better responders: rho = 0.16, p = 0.50; worse responders, rho = -0.08, p = 0.77).

Discussion

We developed a 6-week online, home-based intervention to remediate age-related impairment in sequential decision-making, as well as elucidate individual differences in intervention effects and the underlying brain correlates. The key results can be summarized as follows: intervention-induced effects on value-based sequential decision-making differed between better performing and worse performing older adults; independent of initial performance levels and general cognitive abilities at baseline, older individuals who demonstrated higher hemodynamic response of the DLPFC when performing the task post-intervention also showed a trend towards greater behavioral intervention benefits.

In general, our findings are in line with several previous studies showing that cognitive interventions are effective for improving frontal cognitive functions, in particular, sequential decision-making (e.g., [1] in healthy older adults. The current study, however, revealed inter-individual differences in response to the intervention and provides new insights for elucidating brain correlates dissociating responders from non-responders. Older adults who benefitted from our intervention performed better than baseline from as early as the fifth intervention session. Notably, worse-responding older adults performed equally well at the end of the intervention in the condition demanding complex choice-outcome associations (delayed reward condition) as in the easier condition requiring only concurrent choice-outcome learning (immediate reward condition). This implies that better performing older adults in our sample seemed to have exceeded the level of young adult performance in the same task after the intervention since earlier studies in our group demonstrated that younger adults without intervention showed clear condition effects on performance (cf. [11]. However, it should be noted that this observation is merely descriptive and no direct comparisons were made between trained older adults and untrained younger adults. Nevertheless, these findings suggest that cognitive functions such as decision-making can be trained and improved in older adults by remediating the corresponding neural activity.

These group differences at the behavioral level were accompanied by differential task demand dependent recruitment of DLPFC activity when performing sequential decision-making. Specifically, better responders, but not the worse responders, showed a condition-dependent modulation of DLPFC activity with higher DLPFC recruitment in the delayed reward condition compared to the immediate reward condition at post-intervention. Such an effect was absent in lower performers. Consistent with previous studies on the aging of episodic memory, neural changes were found only in those who benefited from cognitive intervention (e.g., [18]. These findings of a lack of task-related upregulation of DLPFC activation in the more complex condition might have contributed to deficient learning in worse responders. Thus, these results underscore the importance of understanding the neural correlates associated with individual differences in intervention-related gains, so that intervention programs can be individualized to suit the individual needs of older adults.

Amongst the better responders who benefited from the intervention, there was a trend towards a significant correlation between higher DLPFC recruitment in the delayed reward condition at the post-intervention session with practice-related performance gains. However, the findings from the brain-behaviour correlations need to be interpreted with caution as it is only a trend. Although previous research has demonstrated that better responders on a cognitive reserve task had better verbal learning, verbal memory and cognitive flexibility as compared to worse responders [33], there were no significant differences in baseline cognitive measures between better and worse performers in the current study. However, it is important to note that our covariates did not include all of the cognitive processes. Nevertheless, since better responders did not significantly differ from worse responders on baseline task performance and other covariates, the observed differences in practice-related gains between the two performance groups is most likely attributed to differences in condition-dependent DLPFC recruitment. Altogether, this pattern of task-dependent recruitment of cortical activity in better responders resembles that of young adults who showed a higher performance-related DLPFC activity in the delayed compared to the immediate condition [11]. Therefore, age-related decline in sequential decision-making can be remediated by means of cognitive intervention in older adults whose DLPFC activity show task-dependent modulations. However, for worse-responding older adults who do not demonstrate any intervention-related behavioural improvements and corresponding under-recruitment of the DLPFC, it might be beneficial to include other intervention methods such as non-invasive brain stimulation which has the potential to upregulate DLPFC activity and subsequent improvements in behavioural performance [34].

Other than the encouraging behavioral and neural findings in response to a cognitive intervention, our exploratory analyses showed that practice-related gains were maintained at three months after post-intervention, as compared to baseline. In better responders, the decline in performance from post-intervention to 3-months later which seems to be higher than worse-performing individuals could be attributed to two potential factors. Firstly, worse-performing older adults did not show any substantial improvement in their performance, hence, there was not much potential to lose any improvements gained during the intervention. Secondly, skill maintenance might require episodic memory [6], [25], [24] which relies on the medial-temporal brain regions which would have been a more suitable region to correlate with performance at the maintenance phase. However, our current fNIRS montage did not include the medial-temporal brain regions since this was beyond the aims of our current study. Generally, our findings support previous studies which found maintenance of practice-related gains ranging from 3 months up to 10 years in both memory and executive functioning domains [6], [7], [10], [20], [21], [22], [30]. Although it has been suggested that the maintenance of practice-related gains may require booster sessions or adaptive intervention paradigms [5], [9], the current study did not utilize booster sessions or an adaptive paradigm, thus, perhaps an extended amount of intervention is needed to maintain practice-related gains in the absence of booster sessions or adaptive paradigms.

Nevertheless, this study demonstrates that cognitive intervention can result in behavioural and neural improvements even when it is conducted in a home-based setting in the absence of an instructor. One clear limitation of the current study is the small sample size and lack of neural assessment (i.e., DLPFC activity) at baseline, which was unfortunately caused by unavoidable reasons, thus disrupting the plan to measure HbO responses in the lab after baseline behavioral measures were assessed. However, both groups at baseline did not differ in other cognitive tasks requiring DLPFC. Future studies should include multiple sessions of neural assessments throughout the intervention sessions [3] as well as near- and far-transfer tasks to better understand individual differences in intervention-induced brain plasticity and the extent of potential functional impacts of the intervention. Most importantly, however, future studies should also aim to study the effectiveness of non-invasive brain stimulation in combination with cognitive intervention to improve behavioural and neural outcomes in individuals who do not benefit from cognitive intervention.

CRediT authorship contribution statement

Kathleen Kang: Writing – review & editing, Writing – original draft, Visualization, Software, Project administration, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Daria Antonenko: Writing – review & editing, Software, Methodology. Franka Glöckner: Writing – review & editing, Project administration, Methodology. Agnes Flöel: Writing – review & editing, Funding acquisition. Shu-Chen Li: Writing – review & editing, Writing – original draft, Supervision, Resources, Methodology, Funding acquisition, Conceptualization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgments

This project was funded by the German Federal Ministry of Education and Research (Bundesministerium für Bildung und Forschung; FKZ 01GQ1424D) as a subproject in the TRAINSTIM consortium.

Footnotes

Appendix A

Supplementary data to this article can be found online at https://doi.org/10.1016/j.nbas.2024.100109.

Appendix A. Supplementary data

The following are the Supplementary data to this article:

Supplementary data 1
mmc1.docx (4.5MB, docx)

References

  • 1.Antonenko D., Thams F., Grittner U., Uhrich J., Glöckner F., Li S.-C., et al. Randomized trial of cognitive training and brain stimulation in non-demented older adults. Alzheimer’s & Dementia Transl Res Clin Intervent. 2022;8(1):e12262. doi: 10.1002/trc2.12262. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Antonenko D., Thams F., Uhrich J., Dix A., Thurm F., Li S.-C., et al. Effects of a multi-session cognitive training combined with brain stimulation (TrainStim-Cog) on age-associated cognitive decline – study protocol for a randomized controlled phase IIb (Monocenter) trial. Front Aging Neurosci. 2019;11 doi: 10.3389/fnagi.2019.00200. https://www.frontiersin.org/articles/10.3389/fnagi.2019.00200 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Baykara E., Könen T., Unger K., Karbach J. MRI Predictors of Cognitive Training Outcomes. J Cognit Enhancement. 2021;5(2):245–258. doi: 10.1007/s41465-020-00188-y. [DOI] [Google Scholar]
  • 4.Borella E., Carretti B., Cantarella A., Riboldi F., Zavagnin M., De Beni R. Benefits of training visuospatial working memory in young–old and old–old. Dev Psychol. 2014;50(3):714–727. doi: 10.1037/a0034293. [DOI] [PubMed] [Google Scholar]
  • 5.Borella E., Carretti B., Riboldi F., De Beni R. Working memory training in older adults: Evidence of transfer and maintenance effects. Psychol Aging. 2010;25(4):767–778. doi: 10.1037/a0020683. [DOI] [PubMed] [Google Scholar]
  • 6.Brehmer Y., Li S.-C., Straube B., Stoll G., von Oertzen T., Müller V., et al. Comparing memory skill maintenance across the life span: Preservation in adults, increase in children. Psychol Aging. 2008;23:227–238. doi: 10.1037/0882-7974.23.2.227. [DOI] [PubMed] [Google Scholar]
  • 7.Brehmer Y., Rieckmann A., Bellander M., Westerberg H., Fischer H., Bäckman L. Neural correlates of training-related working-memory gains in old age. Neuroimage. 2011;58(4):1110–1120. doi: 10.1016/j.neuroimage.2011.06.079. [DOI] [PubMed] [Google Scholar]
  • 8.Brehmer Y., Westerberg H., Bäckman L. Working-memory training in younger and older adults: Training gains, transfer, and maintenance. Front Hum Neurosci. 2012;6 doi: 10.3389/fnhum.2012.00063. https://www.frontiersin.org/articles/10.3389/fnhum.2012.00063 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Cheng Y., Wu W., Feng W., Wang J., Chen Y., Shen Y., et al. The effects of multi-domain versus single-domain cognitive training in non-demented older people: A randomized controlled trial. BMC Med. 2012;10(1):30. doi: 10.1186/1741-7015-10-30. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Dahlin E., Nyberg L., Bäckman L., Neely A.S. Plasticity of executive functioning in young and older adults: Immediate training gains, transfer, and long-term maintenance. Psychol Aging. 2008;23(4):720–730. doi: 10.1037/a0014296. [DOI] [PubMed] [Google Scholar]
  • 11.Eppinger B., Heekeren H.R., Li S.-C. Age-related prefrontal impairments implicate deficient prediction of future reward in older adults. Neurobiol Aging. 2015;36(8):2380–2390. doi: 10.1016/j.neurobiolaging.2015.04.010. [DOI] [PubMed] [Google Scholar]
  • 12.Iacobucci D., Posavac S.S., Kardes F.R., Schneider M.J., Popovich D.L. The median split: Robust, refined, and revived. J Consum Psychol. 2015;25(4):690–704. doi: 10.1016/j.jcps.2015.06.014. [DOI] [Google Scholar]
  • 13.Karbach J., Könen T., Spengler M. Who Benefits the Most? Individual Differences in the Transfer of Executive Control Training Across the Lifespan. J Cognit Enhancement. 2017;1(4):394–405. doi: 10.1007/s41465-017-0054-z. [DOI] [Google Scholar]
  • 14.Luo, D., Ganesh, S., Koolaard, J., & Luo, M. D. (2018). Package ‘predictmeans’.
  • 15.MacCallum R.C., Zhang S., Preacher K.J., Rucker D.D. On the practice of dichotomization of quantitative variables. Psychol Methods. 2002;7:19–40. doi: 10.1037/1082-989X.7.1.19. [DOI] [PubMed] [Google Scholar]
  • 16.Nagel I.E., Preuschhof C., Li S.-C., Nyberg L., Bäckman L., Lindenberger U., et al. Performance level modulates adult age differences in brain activation during spatial working memory. Proc Natl Acad Sci. 2009;106(52):22552–22557. doi: 10.1073/pnas.0908238106. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Naseer N., Hong K.-S. fNIRS-based brain-computer interfaces: A review. Front Hum Neurosci. 2015;9 doi: 10.3389/fnhum.2015.00003. https://www.frontiersin.org/articles/10.3389/fnhum.2015.00003 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Nyberg L., Sandblom J., Jones S., Neely A.S., Petersson K.M., Ingvar M., et al. Neural correlates of training-related memory improvement in adulthood and aging. Proc Natl Acad Sci. 2003;100(23):13728–13733. doi: 10.1073/pnas.1735487100. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Parisi L., Rocca M.A., Mattioli F., Copetti M., Capra R., Valsasina P., et al. Changes of brain resting state functional connectivity predict the persistence of cognitive rehabilitation effects in patients with multiple sclerosis. Mult Scler J. 2014;20(6):686–694. doi: 10.1177/1352458513505692. [DOI] [PubMed] [Google Scholar]
  • 20.Park, D. C., & Bischof, G. N. (2013). The aging mind: Neuroplasticity in response to cognitive training. Dialogues in Clinical Neuroscience, 15(1), 109–119. 10.31887/DCNS.2013.15.1/dpark. [DOI] [PMC free article] [PubMed]
  • 21.Rebok G.W., Ball K., Guey L.T., Jones R.N., Kim H.-Y., King J.W., et al. Ten-Year Effects of the Advanced Cognitive Training for Independent and Vital Elderly Cognitive Training Trial on Cognition and Everyday Functioning in Older Adults. J Am Geriatr Soc. 2014;62(1):16–24. doi: 10.1111/jgs.12607. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Reijnders J., van Heugten C., van Boxtel M. Cognitive interventions in healthy older adults and people with mild cognitive impairment: A systematic review. Ageing Res Rev. 2013;12(1):263–275. doi: 10.1016/j.arr.2012.07.003. [DOI] [PubMed] [Google Scholar]
  • 23.Roheger M., Meyer J., Kessler J., Kalbe E. Predicting short- and long-term cognitive training success in healthy older adults: Who benefits? Aging Neuropsychol Cogn. 2020;27(3):351–369. doi: 10.1080/13825585.2019.1617396. [DOI] [PubMed] [Google Scholar]
  • 24.Shing Y.L., Werkle-Bergner M., Brehmer Y., Müller V., Li S.C., Lindenberger U. Episodic memory across the lifespan: The contributions of associative and strategic components. Neurosci Biobehav Rev. 2010;34(7):1080–1091. doi: 10.1016/j.neubiorev.2009.11.002. [DOI] [PubMed] [Google Scholar]
  • 25.Shing Y.L., Lindenberger U. The development of episodic memory: Lifespan lessons: episodic memory across the lifespan. Child Dev Perspect. 2011;5(2):148–155. doi: 10.1111/j.1750-8606.2011.00170.x. [DOI] [Google Scholar]
  • 26.Subramaniam K., Luks T.L., Fisher M., Simpson G.V., Nagarajan S., Vinogradov S. Computerized Cognitive Training Restores Neural Activity within the Reality Monitoring Network in Schizophrenia. Neuron. 2012;73(4):842–853. doi: 10.1016/j.neuron.2011.12.024. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Subramaniam K., Luks T.L., Garrett C., Chung C., Fisher M., Nagarajan S., et al. Intensive cognitive training in schizophrenia enhances working memory and associated prefrontal cortical efficiency in a manner that drives long-term functional gains. Neuroimage. 2014;99:281–292. doi: 10.1016/j.neuroimage.2014.05.057. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Tanaka S.C., Doya K., Okada G., Ueda K., Okamoto Y., Yamawaki S. Prediction of immediate and future rewards differentially recruits cortico-basal ganglia loops. Nat Neurosci. 2004;7(8):887–893. doi: 10.1038/nn1279. [DOI] [PubMed] [Google Scholar]
  • 29.von Bastian C.C., Belleville S., Udale R.C., Reinhartz A., Essounni M., Strobach T. Mechanisms underlying training-induced cognitive change. Nature Reviews. Psychology. 2022;1(1):Article 1. doi: 10.1038/s44159-021-00001-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Willis S.L., Tennstedt S.L., Marsiske M., Ball K., Elias J., Koepke K.M., Morris J.N., Rebok G.W., Unverzagt F.W., Stoddard A.M., Wright E., ACTIVE Study Group, for the Long-term effects of cognitive training on everyday functional outcomes in older adults. J Am Med Assoc. 2006;296(23):2805–2814. doi: 10.1001/jama.296.23.2805. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Xu, Y., Graber, H. L., & Barbour, R. L. (2014). nirsLAB: A Computing Environment for fNIRS Neuroimaging Data Analysis. Biomedical Optics 2014 (2014), Paper BM3A.1, BM3A.1. 10.1364/BIOMED.2014.BM3A.1.
  • 32.Zelinski E.M., Spina L.M., Yaffe K., Ruff R., Kennison R.F., Mahncke H.W., et al. Improvement in Memory with Plasticity-Based Adaptive Cognitive Training: Results of the 3-Month Follow-Up. J Am Geriatr Soc. 2011;59(2):258–265. doi: 10.1111/j.1532-5415.2010.03277.x. [DOI] [PubMed] [Google Scholar]
  • 33.Zihl J., Sousa N., Walther K., Fink T., Schmid A., Almeida O.F.X. Cognitive reserve in the healthy elderly: Cognitive and psychological factors. ScienceOpen Res. 2015 10.14293/S2199-1006.1.SOR-SOCSCI.A5KKMA.v2. [Google Scholar]
  • 34.Schommartz I., Dix A., Passow S., Li S.-C. Functional Effects of Bilateral Dorsolateral Prefrontal Cortex Modulation During Sequential Decision-Making: A Functional Near-Infrared Spectroscopy Study With Offline Transcranial Direct Current Stimulation. Frontiers in Human Neuroscience. 2021;14 doi: 10.3389/fnhum.2020.605190. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary data 1
mmc1.docx (4.5MB, docx)

Articles from Aging Brain are provided here courtesy of Elsevier

RESOURCES