Alzheimer’s disease (AD) is a representative neurodegenerative disorder characterized by the deposition of amyloid beta (Aβ) and tau proteins, leading to impairment of cognitive function and the ability to perform daily activities (Scheltens et al., 2016). Mild cognitive impairment (MCI) is a prodromal stage of AD and approximately 10–15% of patients with MCI have been reported to convert to dementia each year (Gauthier et al., 2006). Despite the high risk of progression to dementia, treatment options for MCI are limited. Although several drugs have been investigated in clinical trials, their efficacy in delaying or preventing progression to dementia is modest and uncertain (Anderson et al., 2017). Therapeutic approaches such as cognitive intervention (Jean et al., 2010), physical exercise (Lautenschlager et al., 2010), and dietary modification have demonstrated some promising outcomes in terms of changes in cognitive function and biomarkers (Singh et al., 2014). However, further research is required to establish these interventions as effective strategies for preventing AD. Additionally, the complexity of these interventions can make them challenging for individuals with MCI to implement and maintain consistently (Coley et al., 2019), highlighting the need for a simple, fixed intervention that can be sustained over a specific period. In this regard, there is currently a need for alternative interventions that are accessible to individuals with MCI and can improve cognitive function and mitigate neurodegenerative changes. An interest in non-invasive brain stimulation treatment methods is also increasing, and in particular, transcranial direct current stimulation (tDCS), which is highly accessible in terms of cost and portability, is proposed as one of the appropriate treatment options for MCI (Liu et al., 2017). tDCS modulates the excitability of cortical neurons depending on the current flow direction and has synaptic after-effects through long-term potentiation (LTP), affecting neuroplasticity (Nitsche et al., 2003). It is also suggested to normalize brain function in patients with AD and facilitate the clearance of Aβ by modulating the integrity of the brain–blood barrier (Meinzer et al., 2015; Shin et al., 2020). Previous studies have shown that cognitive function can be improved through single or multi-session tDCS in patients with AD and MCI (Lu et al., 2019; Manenti et al., 2020). Moreover, long-term application of this technique is expected to modulate disease progression (Im et al., 2019; Yang et al., 2019).

Among intermediate phenotypes related to AD, changes in white matter microstructure have been demonstrated to occur earlier in dementia than changes in brain function, making it a useful early warning sign for the development of the disease (Zhang et al., 2009). Additionally, advanced imaging techniques, such as diffusion tensor imaging (DTI), can be used to quantify white matter assessment, allowing for more precise and objective measurement of changes over time (Le Bihan and Johansen-Berg, 2012). Patients with MCI have shown a reduced fractional anisotropy (FA) value, which reflects white matter integrity, in multiple posterior white matter regions, as well as in frontal, temporal, parietal, and occipital white matter and association fibers compared to normal subjects (Medina et al., 2006). Moreover, lower FA values of the splenium of corpus callosum and crus of fornix have accurately differentiated between amnestic MCI patients and control subjects (Medina et al., 2006). Furthermore, the deterioration of white matter microstructure has predicted a more rapid transition to MCI in cognitively intact older adults (Shafer et al., 2022). Additionally, prior research has indicated that the white matter microstructure changes in a non-linear manner as the clinical phenotype progresses and the extent of Aβ deposition increases (Dong et al., 2020; Pereira et al., 2021). In this regard, a connection to compensatory mechanisms has been proposed (Raghavan et al., 2021).

Despite the clinical evidence of white matter microstructure as an intermediate phenotype in the prodromal AD phase, few studies have examined changes in white matter microstructure after tDCS in MCI patients. In other vascular and neurodegenerative diseases, previous studies have shown that FA values increase in the frontal lobe after tDCS application in stroke patients with memory impairment (Hua et al., 2022). The white matter integrity of ventral language pathways has also predicted letter accuracy in primary progressive aphasia after tDCS application (Zhao et al., 2021).

Precision medicine, which involves personalized strategies based on an individual’s genetic, biomarker, and clinical profile, has gained traction as a promising approach in the management of AD (Isaacson et al., 2018). It can help optimize tDCS treatment response and lead to more effective and efficient AD management by considering individual AD risk factors. Aβ deposition has been shown to affect the rate of cognitive decline and intermediate phenotype of brain structure and function, as well as the risk of transitioning to dementia in patients with MCI (Lukiw et al., 2020). In addition, APOE ε4 allele has been shown to affect the likelihood and severity of the Aβ pathophysiological cascade and is responsible for the greatest proportion of genetic risk factors for sporadic AD (Harrison et al., 2020). Brain-derived neurotrophic factor (BDNF) is a protein that impacts the growth and maintenance of neurons and has been known to play an important role in LTP-like neuroplasticity induced by tDCS (Chaieb et al., 2014). Previous studies have shown that its levels are decreased in AD (Ng et al., 2019). Moreover, research has suggested that BDNF Val66Met polymorphisms may modulate the risk of AD by affecting BDNF levels (Du et al., 2018). In regard to sex, research suggests that there are distinct differences in the clinical and intermediate phenotype patterns of males and females with AD (Ferretti et al., 2018). Specifically, studies have shown that females with MCI and dementia tend to experience a more rapid cognitive decline and brain atrophy after diagnosis (Ferretti et al., 2018). However, there have been few studies examining the impact of tDCS treatment based on individual AD risk factors, which would allow for more precise medical treatment strategies using tDCS in the prodromal stage of AD. In our previous pilot study in MCI patients, we have found that brain functional integration and segregation parameters differ after sequential tDCS to the left dorsolateral prefrontal cortex (DLPFC) depending on Aβ deposition and APOE ε4 carrier status (Kang et al., 2021). The BDNF Val66Met polymorphism has been reported to have a duration-dependent effect on tDCS-induced motor cortex plasticity in older adults (Puri et al., 2015). However, other studies have exhibited inconsistent results on the impact of BDNF Val66Met polymorphism on cortical excitability (Antal et al., 2010; Teo et al., 2014). Regarding sex, previous research has reported a differential impact of tDCS on practice-related executive function in older adults, with higher current density observed in female older adults (Fehring et al., 2021). Therefore, it is worthwhile to investigate how the impact of tDCS on white matter microstructure integrity in the AD prodromal phase is modified by representative individual AD risk factors as effect modifiers.

In this context, we aimed to evaluate whether a 2-week application of sequential tDCS alters white matter microstructure integrity in patients with MCI and whether it depends on effect modifiers consisting of individual risk factors for AD. Furthermore, for white matter microstructure tracts that exhibit a significant association with baseline Aβ deposition, we assessed the correlation with the differences in cognitive function scores to elucidate the nature of the alteration.

The present study was designed to investigate how the effect of 2 weeks of tDCS on white matter microstructural integrity in MCI patients varies according to an effect modifier composed of individual factors of AD, including Aβ deposition, APOE ε4 carrier status, BDNF Val66Met polymorphism status, and sex.

This study found a significant interaction between AD risk factors and tDCS for white matter microstructural integrity in tracts at risk for AD. The interaction between APOE ε4 carrier status and tDCS was attributed to increased FA in Rt. UF of APOE ε4 carriers after sequential tDCS sessions. The UF is a white matter tract involved in cognitive functions such as episodic memory and language (Von Der Heide et al., 2013), and previous studies have shown that FA levels in the UF decrease with the progression of AD (Liu et al., 2011; Qin et al., 2021). APOE ε4 carrier status has been found to differentially affect the integrity of white matter microstructure depending on age and severity of AD (Kljajevic et al., 2014), inducing better integrity of the UF compared with APOE ε4 non-carrier without a family history of AD (Adluru et al., 2014). However, there is a paucity of research on the effects of tDCS on white matter integrity at any stage of AD, including MCI, and there is also a lack of research on potential interactions with effect modifiers such as APOE ε4 carrier status. Taken together, the increased white matter microstructural integrity of the UF observed in APOE ε4 carriers after 10 sessions of sequential tDCS may reflect a compensatory mechanism for AD progression. However, given that white matter integrity is an intermediate phenotype for AD progression, further work is needed to clarify that changes in this metric are compensatory for the cognitive and functional deterioration. Given that the transition from MCI to AD takes 4–5 years (Vermunt et al., 2019), it is important to assess the impact of long-term changes in white matter microstructural integrity on changes in cognitive function and the risk of transition from MCI to AD.

In the present study, sequential tDCS led to increased FA in the Rt. UF of female subjects due to the sex*tDCS interaction. This may be due to the fact that white matter microstructure deterioration in females begins later and progresses more gradually than in males (Toschi et al., 2020), and older females receive a higher intensity of tDCS current at the target site compared to males, which may be due to the age-dependent sex difference in tDCS current intensity resulting from cerebral atrophy (Bhattacharjee et al., 2022).

The BDNF Val66Met polymorphism status of participants in the research interacted with the tDCS application, resulting in increased MD in the Rt. UF and decreased MD and RD in the Lt. cingulum of the cingulate cortex of the Met non-carriers after sequential tDCS. The BDNF has been shown to affect white matter microstructure by modulating myelinogenesis (Du et al., 2003), but its impact varies depending on factors such as age and type of tract (Voineskos et al., 2011; Tost et al., 2013). A previous study found that the BDNF Val66Met polymorphism interacts with age to affect white matter microstructure, particularly in corticocortical association tracts and late-myelinating fiber tracts in the brain, suggesting that older Val66 homozygotes may be more susceptible to changes in white matter microstructure (Voineskos et al., 2011). However, in the present study, the Rt. UF and Lt. cingulum of the cingulate cortex showed differential changes after tDCS application in Val66 homozygotes, despite both being the latest-myelinated fibers. Furthermore, changes in the microstructural integrity of the Rt. UF were found to be different between APOE ε4 carriers and older Val66 homozygotes, although both groups are susceptible to AD. Relative to APOE ε4 carrier status, the BDNF Val66Met polymorphism has shown inconsistent effects on white matter integrity, with both increased and decreased FA and RD reported in different brain tracts (Soliman et al., 2010; Chiang et al., 2011; Tost et al., 2013). This highlights the complexity of the interaction between BDNF Val66Met polymorphisms, white matter integrity, and the response to tDCS. Further research is therefore needed into the factors that may contribute to this inconsistency.

This study also found that tDCS interacts with multiple factors to affect white matter integrity who were Val66 homozygotes and APOE ε4 carriers following sequential tDCS. Similarly, female with Val66 homozygosity showed a significant improvement in white matter integrity in the Rt. UF after tDCS sessions. Previous studies have shown accelerated Aβ deposition in APOE ε4 carriers with BDNF Met carriers (Stonnington et al., 2020; Riphagen et al., 2022), but there is limited investigation into the impact on the white matter integrity. Regarding the sex, although some studies suggest a sexually dimorphic effect of the BDNF Met66 allele on AD susceptibility (Fukumoto et al., 2010), others contradict this (Voineskos et al., 2011). In the prodromal phase of AD, Val66 homozygotes with a higher susceptibility to age-related decline in white matter integrity (Voineskos et al., 2011) are assumed to induce a greater increase in integrity in the AD susceptible tract in response to sequential tDCS in the presence of AD risk factors. In addition, the UF that shows an interaction between tDCS and AD risk factors in this study is a pathway that connects the anterior temporal lobe to the orbitofrontal cortex and is involved in cognitive functions such as episodic memory and language (Von Der Heide et al., 2013). It is therefore worth investigating how improved integrity in this pathway affects the clinical course of AD in the long term.

In this investigation, we observed a significant correlation between baseline Aβ deposition and enhanced FA in the left cingulum within the hippocampal region, following the administration of sequential tDCS. This aligns with prior research that reported diminished fiber connections in this specific area among patients in the early stages of AD and those with MCI (Zhou et al., 2008). Moreover, we identified a positive association between the increase in FA of the Lt. cingulum in the hippocampal region and improvements in delayed memory performance, corroborating earlier findings (Zhou et al., 2008). In subgroups stratified by AD risk factors, we also found a positive association between Aβ deposition and the difference in FA of the left cingulum in the hippocampal area in APOE ε4 carriers. However, regarding the FA of this tract, there was not significant relationship with the differences in neuropsychological performances. Taken together, these observations suggest a potential compensatory mechanism whereby Aβ deposition triggers alterations in the white matter integrity of the Lt. cingulum in the hippocampal area, especially in the APOE ε4 carriers. This change could be catalyzed by the application of sequential tDCS during the prodromal phase of AD. Nonetheless, further research with larger sample size is warranted to substantiate these preliminary insights.

A limitation of this study is the duration of sequential tDCS. As this study only involved a 10-session application of tDCS, conducting a long-term study would offer a more comprehensive understanding of the clinical implications of the compensatory changes in white matter microstructure integrity that were inferred from this study. Furthermore, conducting a study that includes a sham stimulation group would enhance our understanding of the role and clinical implications of the effect modifiers identified in the present study with MCI patients.

The purpose of this research was to assess the impact of a 2-week application of sequential tDCS on changes in white-matter microstructure integrity of MCI patients, taking into account individual factors for AD. Additionally, we examined the association between baseline Aβ deposition and changes in white matter microstructure integrity, as well as the relationship between changes in cognitive function and those in white matter integrity. The study revealed a significant effect of the effect modifiers*tDCS interaction on the white matter microstructure integrity of the AD vulnerable tract. Furthermore, there was a positive association between Aβ deposits and changes in the integrity of the white matter tract which reflects the AD progression. Finally, the changes in the tract of interest also exhibited a positive relationship with the differences in the memory performance in the prodromal stage of AD. AD is a multi-faceted neurodegenerative disease that elicits various responses to treatments among patients diagnosed with MCI. Therefore, it is essential to apply a precision medicine therapeutic approach that takes into account individual AD-related factors, especially regarding non-invasive brain stimulation. In this regard, this study provides a cornerstone for the clinical importance of precision medicine approaches in tDCS and in the field of non-invasive brain stimulation therapy for patients on the AD trajectory. Additionally, conducting additional research to address the limitations of the current study would provide an opportunity to reassess the therapeutic potential of tDCS in the prodromal AD phase, which has limited treatment options available.

The datasets presented in this article are not readily available because the datasets generated or analyzed during the current study are not publicly available due to the Patient Data Management Protocol of Yeouido Saint Mary’s Hospital but are available from the corresponding author upon reasonable request. Requests to access the datasets should be directed to DWK, kato7@hanmail.net.

The studies involving humans were approved by Institutional Review Board of the Catholic University of Korea. The studies were conducted in accordance with the local legislation and institutional requirements. Written informed consent for participation in this study was provided by the participants’ legal guardians/next of kin.

DWK: conceptualization, methodology, data curation, writing–original draft, visualization, formal analysis, and funding acquisition. S-MW: methodology, data curation, and writing–review and editing. YU: software and investigation. SK: methodology and data curation. TK: methodology and data curation. DK: methodology and data curation. CL: conceptualization and supervision. HL: conceptualization, methodology, writing–review and editing, supervision, and project administration. All authors contributed to the article and approved the submitted version.

This research was supported by the National Research Foundation of Korea (NRF) grant funded by the Korean government (Ministry of Science and ICT; No. 2019R1C1C1007608) and a grant of the Korea Health Technology R&D Project through the Korea Health Industry Development Institute (KHIDI), funded by the Ministry of Health & Welfare, Republic of Korea (grant number: HI22C0467).

HL, TK, and DK were employed by NEUROPHET Inc.

The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. The data processing services provided by NEUROPHET Inc. were utilized to enhance the quality and analysis of the brain imaging data collected during the study. Authors declare that the research outcomes and conclusions remain unbiased and are not influenced by any commercial interests associated with the NEUROPHET Inc.’s products or services.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.