Progressive difficulty in performing everyday functional activities is a key diagnostic feature of dementia. Yet not much is known about the neural underpinnings of functional decline, particularly during the very early stages of dementia. Early intervention before impairment in everyday functions is observed offers the best hope of reducing the burdens of Alzheimer disease (AD) and other dementias. However, to justify early intervention, those at risk need to be detected earlier and more accurately.

The decline in complex daily function (CdF) such as managing medications and finances has been reported to precede impairment in basic activities of daily living (eg, eating and dressing). Our goal is to establish the neural signature of decline in CdF during the preclinical dementia period. We recently reported that limitations in CdF were more prevalent or declined faster in those who converted to mild cognitive impairment (MCI) [1]. We compared CdF profiles at baseline in 59 community-dwelling older individuals with normal cognitive performance who went on to develop incident MCI (“pre-MCI”) with 284 older individuals who remained cognitively normal over follow-up. While limitations in CdF were reported to be subtle, the mean number of limitations at baseline was 3.1 (SD 3.0) in the pre-MCI cases and 2.0 (SD 2.4) in normal controls (P=.003). This is in line with studies using amyloid-β (Aβ), phosphorylated τ, neurofilament (also referred to as AT(N) classification system) [2] to stratify dementia risk in cognitively normal older adults, showing the presence of AT(N) biomarkers to be associated with more limitations in CdF [3,4].

While combining behavioral (eg, cognition and CdF assessment) with biological (eg, AT(N)) markers has further improved early risk assessment, it is insufficient to justify intervention during the preclinical dementia stages. For example, up to one-third of older individuals with significant accumulation of AD-related and vascular pathology will not progress to MCI and dementia stages [5,6]. In this proposal, we seek to further improve risk assessment by establishing the neural signature of CdF. We posit that abnormal patterns of neurophysiological activation precedes and can serve as a robust predictor of change in CdF, which in turn increases the risk of conversion to dementia [7]. To test this hypothesis, we designed complex gait tasks. Gait has physical and cognitive contributions. We showed that gait performance (eg, gait speed) during normal and dual-task conditions is associated with scores on the Activities of Daily Living–Prevention Instrument (ADL-PI) [8], which includes questions about abilities to manage money, drive, do the laundry, and use appliances [9,10]. We applied a novel electroencephalographic-based mobile brain-body imaging approach to measure gait-related brain activation while participants perform complex gait-based functional tasks [11-19]. Raising the complexity of the gait task may allow early distinction of individuals with and without the risk of developing limitations in CdF by increasing the cognitive demands of the task. We designed simple (walking-only) and complex (dual-task walking) gait-based tasks to measure differential activation (complex minus simple gait). This proposal builds on and extends 2 decades of work by our team showing that changes in gait (eg, slowing of walking speed) during simple and complex walking conditions are robust predictors of decline in CdF, cognitive decline, and progression to dementia [9,10,20-30].

Our overarching hypothesis illustrated in Figure 1 is that the presence of dementia-related pathology during the preclinical period of dementia is associated with a unique gait-related electroencephalographic (grEEG) pattern that predicts a subsequent decline in CdF. The first aim of this study is to establish the neural signature of CdF. Our preliminary findings show that individuals reporting CdF limitations can be characterized by weaker pre as well as postcentral gyrus and stronger fronto-medial activation during complex gait (eg, dual-task walking). This unique grEEG activation pattern (neural signature) will predict CdF decline longitudinally. Our second aim is to determine the association of dementia-related pathology with the incidence of the neural signature of CdF in individuals without the neural signature at baseline. Specifically, we hypothesize that plasma-based assays of Aβ, T, and white matter hyperintensities (WMH) at baseline will independently and in combination predict the incidence of the neural signature of CdF. We also need to understand if and how behaviors are known to increase dementia risk (eg, physical inactivity) and may modify the neural signature of CdF as a prelude to developing interventions. Our third aim is to determine whether physical inactivity is associated with the prevalence and incidence of the neural signature of CdF in our sample.

This project is part of a larger effort at the Division of Cognitive and Motor Aging at Albert Einstein College of Medicine (AECOM) to go beyond memory-centered depictions of AD and consider cognitive processes related to motor and sensory abilities as important additional characterization of AD and related dementias. We are partnered with an ongoing investigation (R01AG075679-01) that seeks to establish visual-somatosensory integration (VSI study) as a novel marker of preclinical AD. With Aβ accumulation occurring early in sensory association areas [31], the VSI study (principal investigator [PI] JRM) seeks to define multisensory integration as a preclinical marker of AD and related dementias (see [32] for more details). The VSI study serves as a parent study for cross-enrollment and includes older adults across the cognitive spectrum from cognitively normal to MCI. Participants in the VSI study with confirmed preclinical AD diagnosis (defined as Aβ positive) will undergo positron emission tomography testing. Collectively, these studies will broaden the scope of our investigations in synergistic ways to deepen our understanding of the very early preclinical dementia period.

We propose a prospective observational cohort study of 180 older adults who are cognitively unimpaired. Figure 2 shows the recruitment and timeline of procedures. We will cross-enroll participants from an active study (R01AG075679, PI JRM). We will use telephone-based screening procedures to enroll and follow a community-based cohort. Potential participants aged 65 year and older from the greater New York City area are first contacted by mail and then by telephone to explain the purpose and nature of this study. The telephone interview includes verbal consent, a brief medical history questionnaire, and 2 brief cognitive screens [33,34]. We will ensure that enrollment reflects the race and ethnic diversity of our catchment area. Following the interview, potential participants are invited for further evaluation at our research center. If eligible, written informed consents are obtained at study visits as per institutional review board (IRB) guidelines. This study is approved by the AECOM IRB.

Table 1 lists instruments used to assess function and other domains. The ADL-PI will serve as our primary outcome measure. It is a 15-item questionnaire that was developed to target the earliest changes in complex daily activities [8]. A performance-based instrument, the Everyday Problems Test [35], will serve as our secondary CdF outcome measure. It is a 42-item paper-pencil test that examines abilities to understand and execute complex daily activities (eg, write a check).

Neuroimaging and plasma-based blood testing will be conducted during baseline visits for each participant. C2N Diagnostic will quantify plasma τ217 (nonphosphorylated and phosphorylated) and plasma amyloid (Aβ42, Aβ40), and provide plasma ApoE prototyping using liquid chromatography-tandem mass spectrometry platforms (PrecivityAD) [53-55]. Blood samples will be collected and stored in the biorepository at AECOM. Physical activity will be assessed at baseline and during annual in-person visits using the Physical Activity Scale for the Elderly [36] as well as objectively using accelerometry [56] over 7 days (AX3; Axivity Ltd). The Physical Activity Scale for the Elderly [57,58] and AX3 [59] are reliable and valid measures of physical activity.

Established clinical consensus case conference procedures [60], where participants’ demographic, neuropsychological, neurological, psychosocial, and functional test results are reviewed by a multidisciplinary clinical team consisting of neurologists and neuropsychologists, will be used to determine normal cognitive status. Cognitively unimpaired individuals testing positive for Aβ are classified as preclinical AD [2]. Cognitive status—normal, MCI, and dementia—will be assigned at baseline and during yearly follow-up visits. Individuals with MCI and dementia at baseline are excluded from this project.

For aim 1, the primary outcome is complex daily functional limitations measured with the ADL–PI, as reported in our previous publication [8]. For aims 2 and 3, the outcome is the incidence of the neural signature of CdF (see section below for definition). All outcome measures are assessed annually over a maximal period of 4 years.

Protocols for the proposed research project are approved by the AECOM’s IRB (2023-14773). Before enrollment, written consent is obtained from all interested persons who meet the inclusion criteria at the time of the visit. Confidentiality will be preserved by the use of unique ID codes for identification. ID and name associations will be password-protected in an encrypted master file to which only the PI and study coordinator have access. Participant data, including computer data disks, will be kept in a locked room. Participants will receive monetary compensation (US $20 per hour), free lunch, and transportation for participating in this study.

We designed simple (walking-only) and complex (dual-task walking) gait tasks to measure differential activation (complex minus simple gait). We provide preliminary findings showing that individuals reporting CdF limitations can be characterized by a unique grEEG pattern: weaker sensorimotor activation, measured by pre or postcentral gyrus 13-28 Hz desynchronization, and stronger motor control activation, measured by fronto-medial 3-7 Hz synchronization. Figure 3 shows an individual performing the visual-perturbed walking task. The participant walking on a treadmill wearing an electroencephalographic cap is immersed into a large-scale visual flow field designed to destabilize posture [61]. We superimpose sideway shifts onto a display of dots radiating outward from a central point of extension to create visual perturbations. We validated the task in prior studies [12,15]. Participants wore harnesses for safety.

In our first attempt to identify the neural signature of CdF, we stratified individuals by their cognitive performance [19]. Cognitively normal individuals (Montreal Cognitive Assessment [MoCA] ≥ 22 [62,63]) were dichotomized into low (score of 22-26; n=10) and high (score 27+; n=16) performing groups. We discovered a unique grEEG activation pattern time-locked to the gait cycle in low-cognitive performers: weaker central gyrus β (13-28 Hz) desynchronization paired with stronger fronto-medial θ (3-7 Hz) synchronization during visually perturbed compared to unperturbed walking. We also found correlations between θ synchronization and worse MoCA scores and between β desynchronization and better MoCA scores. Amplitude suppression (ie, desynchronization) of brain oscillatory activity within 8 to 28 Hz over pre as well as postcentral gyrus during movement is considered a neurophysiological marker of sensorimotor activation related to preparation and execution of movement [64-70]. We suggest that posterior parietal and sensorimotor activation during gait adjustment is related to prioritizing inputs from visual, somatosensory, and vestibular systems based on reliability to build accurate proprioception subserving motor output [19]. On the other hand, increased θ activity (ie, synchronization) over the fronto-medial cortex is thought of as a mechanism by which the need for cognitive or motor control is realized and signaled across brain regions [71,72]. In the context of movement, we and others observed fronto-medial θ in moments of postural instability [12,15,73-78].

To test whether this unique grEEG activation pattern also associates with the limitation in CdF, we dichotomized our sample in the following way: (1) compute the difference in activation between unperturbed and perturbed walking; (2) use inverse source localization techniques to estimate intracranial sources of scalp-recorded activity; (3) separate individuals with source localized to fronto-medial regions into groups of weak and strong θ synchronizers using the median; (4) separate individuals without sources in fronto-medial and with sources in sensorimotor region into groups of weak and strong β desynchronizers using the median (we chose right sensorimotor β because effects where strongest for that region and frequency: P=.04) [19].

Figure 4 shows preliminary findings in support of our hypothesis: Individuals characterized by a unique grEEG pattern report limitations in CdF to a higher degree.

A major effort aimed at curbing the rapid rise of dementia revolves around early intervention. To substantiate early intervention, it is crucial to enhance the accuracy of early detection. We focus on the neural signature of early limitations in CdF to improve detection. Walking is central to many instrumental and complex activities [9,10]. To determine the neural signature of CdF, our team [11-19] and others [73,87,98-104] have established the validity and reliability of neurophysiological recordings while participants are in motion. We designed novel grEEG protocols for this proposal and selected walking as an activity that, compared to other daily activities (eg, preparing a meal), allows for the use of standardized, reproducible, reliable, and well-tolerated study protocols across participants [30,39,105,106].

Our preliminary findings in 28 older adults without MCI/dementia show that reporting of CdF limitations is associated with a unique grEEG activation pattern: 13-28 Hz central gyrus desynchronization paired with 3-7 Hz fronto-medial synchronization. We term this electroencephalographic activation pattern the neural signature of CdF. This subsample also had higher Aβ burden, smaller brain volume, and WMH in regions affected early by dementia and engaged in less physical exercise.

Some limitations are important to note. The ceiling effect with a majority reporting no limitations in CdF is a concern (eg, in Marshall et al [79] study only 16% reported limitations). However, in our study, 30% of individuals reported limitations [1]. To optimize detection, we choose 2 validated and reliable instruments (self-report [107] and paper-pencil test [35] with numerous dependent variables such as self-assessment, time-to-complete, and number and type of errors to quantify performance) covering a broad range of complex daily activities. Further, different processes denoted by Aβ deposition and WMH and their respective roles in the etiology of AD and other dementias are matters of ongoing debates. The prevalence of amyloid deposition in normal aging warrants consideration of Aβ in concert with phosphorylated τ and neurofilament light assays to recognize the multitude and temporal order of pathological processes leading up to clinical signs.

In summary, establishing a noninvasive neurophysiological signature of prevalence and incidence of CdF decline will refine pre-MCI stage characterization. As our recent study shows [1], both individuals who progress and do not progress to MCI may report mild limitations in CdF during the preclinical period, though more limitations are reported in the individuals who go on to develop MCI. By establishing the clinical relevance and biological basis of the neural signature of CdF decline, we aim to improve prediction during preclinical stages of ADs and other dementias.