The Alzheimer’s Pathways Sleep Study (ALPS): an experimental randomized controlled trial to improve cognition and Alzheimer’s pathophysiology through slow-wave sleep

Abstract

Background

Slow-wave activity (0.5–4 Hz electroencephalographic activity) during non-rapid eye movement sleep is consistently associated with better cognitive performance in older adults. Slow-wave activity is known to regulate synaptic plasticity and may thereby mitigate excitotoxicity and accumulation of Alzheimer’s pathology. Paradoxically, longer total sleep times in older adults are often associated with poorer cognition and general health. Conventional behavioral sleep treatments robustly increase sleep efficiency and sleep time, but do not consistently enhance slow-wave activity and have shown only subtle effects on cognition. Thus, enhancement of slow-wave activity may be a critical target for sleep-based cognitive enhancement. The Alzheimer’s Pathways Sleep Study (ALPS) uses a novel time-in-bed (TiB) restriction intervention designed to increase slow-wave activity through homeostatic sleep drive and assesses improvements in measures of excitotoxic hippocampal hyperactivation, plasma levels of amyloid beta (Aβ), and overnight memory retention.

Methods

ALPS is a randomized controlled trial designed to increase slow-wave activity behaviorally in older adults with poor sleep. Target enrollment is 116 participants aged 65–85. Participants are randomized to a TiB restriction intervention or an attention-matched control intervention. Participants randomized to TiB restriction follow a sleep schedule restricting their TiB to 85% of their habitual TiB for 4 weeks. The control group follows their habitual TiB for 4 weeks. Both groups are monitored with diary, actigraphy, check-in calls, and sleepiness ratings over the 4 weeks. The primary outcomes for four specific aims are (1) absolute power in the slow oscillation (0.5–1 Hz) range during non-rapid eye movement sleep, (2) hippocampal activation during memory encoding, (3) plasma Aβ_1–42, and (4) overnight word pair memory retention. Here we describe how the ALPS intervention is administered, the protocols and scripts implemented to maximize adherence and safety, and outcomes measured to test the proposed conceptual model linking slow-wave activity with excitotoxic hyperactivation, Alzheimer’s pathophysiology, and memory performance.

Discussion

Behavioral enhancement of homeostatic sleep drive through TiB restriction is a promising approach to improve memory performance and Alzheimer’s pathophysiology. Safety measures, as described here, should be implemented to minimize risks associated with TiB restriction.

Trial registration

ClinicalTrials.gov NCT05138848. Registered on December 1, 2021.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13063-026-09505-w.

Introduction

Background and rationale

Sleep is involved in regulating various aspects of brain health, cognition, and mechanisms of Alzheimer’s disease in older adults [1, 2]. Pharmacological treatments to promote sleep are known to cause significant cognitive impairments and hangover effects [3]. Thus, the design of non-pharmacological sleep interventions that modify cognition and disease pathology is a top priority for the field of Alzheimer’s disease [4–6]. Nonetheless, there is a lack of clinical trials demonstrating that improving habitual sleep in older adults leads to improvements in cognition and Alzheimer’s disease pathophysiology [7–10].

Multiple aspects of sleep that decline with aging are commonly associated with better cognitive performance [11]. For instance, greater sleep efficiency (the amount of time in bed spent sleeping) is often associated with better memory recall and executive functions in older adults [7, 8, 12] and is commonly disrupted in Alzheimer’s disease [13–18]. Similar patterns are found with sleep oscillations such as slow-wave activity [7, 8, 14], which is characterized by 0.5–4 Hz electroencephalographic activity during non-rapid eye movement (NREM) sleep. Paradoxically, numerous studies and meta-analyses show that very long total sleep times (i.e., ≥9 h) are associated with poorer cognition and general health [11, 19–21]. Further, increased sleep duration is longitudinally associated with cognitive decline. Thus, it is not merely the amount of sleep, but rather more consolidated, deeper sleep (greater sleep efficiency and slow-wave activity) that appears to be important for maintaining cognition in older adulthood [22].

The most direct behavioral route to increasing slow-wave activity is through increasing homeostatic sleep drive (referred to as “Process S”) by prolonging time spent awake [23, 24]. Conventional behavioral sleep interventions commonly use time-in-bed (TiB) restriction to increase homeostatic sleep drive and reduce the amount of TiB spent awake to associate the sleep environment with sleep [25]. For example, in Brief Behavioral Treatment for Insomnia, TiB is typically reduced to the number of sleep hours + 30 min initially and then gradually increased as insomnia symptoms subside [25, 26].

By maximizing homeostatic sleep drive through TiB restriction, we may reduce not only the amount of time in bed spent awake, but also the amount of time spent in lighter sleep (i.e., stage N1), while prioritizing deeper, more restorative aspects of sleep (i.e., slow-wave activity) [22]. Moreover, slow-wave activity increases may persist with prolonged TiB restriction, assuming moderate levels of total sleep time are met. In a pilot study in older adults with poor sleep, we found that 1 week of TiB restriction to 75% of habitual TiB led to increased slow-wave activity, sleep efficiency, and reduced nighttime wakefulness. Although total sleep time significantly decreased, paired associate learning improved in those undergoing TiB restriction, but not in an attention-matched control group [22].

The Alzheimer’s Pathways Sleep Study (ALPS) is based on the premise that there may be a “sweet spot” for TiB for older adults such that slow-wave activity is maximized while maintaining adequate total sleep time. ALPS examines slow-wave activity and its role in synaptic downscaling [27] as a mechanism of reducing excitotoxicity and risk for Alzheimer’s disease in older adults with poor sleep. Considering that sleep recommendations are commonly based on number of hours [28], TiB restriction may be a counterintuitive approach to improve cognition through sleep. However, TiB restriction is a simple tactic that is easily implemented to improve aspects of sleep most reliably implicated in cognitive aging.

Objectives

The ALPS study extends the pilot study [22] TiB restriction intervention by examining effects of TiB restriction over 4 weeks on SWA, functional imaging markers of excitotoxic hippocampal hyperactivation, plasma Alzheimer’s disease biomarkers, and overnight memory retention. The protocol involves TiB restriction to 85% of habitual TiB. TiB restriction to 85% TiB was selected rather than 75% as with the pilot study, to minimize daytime sleepiness and maximize tolerability over 4 weeks [22]. Four specific aims test a comprehensive conceptual model with the central hypothesis that enhanced SWA will reduce excitotoxic hippocampal hyperactivation and improve plasma Aβ levels and overnight memory retention. The specific aims for ALPS are:

• Aim 1: Examine whether TiB restriction increases SWA after 4 weeks.

• Aim 2: Examine whether TiB restriction improves fMRI indices of excitotoxic hyperactivation.

• Aim 3: Examine whether TiB restriction improves plasma Aβ levels.

• Aim 4: Examine whether TiB restriction is associated with improved memory retention.

The current report describes (1) the ALPS TiB restriction intervention and how it is administered, (2) the exclusion criteria and monitoring protocols to maximize safety in the context of TiB restriction, and (3) the study measures to test effects on cognitive and Alzheimer’s pathophysiology measures.

Methods

Trial design

ALPS is a randomized controlled trial (RCT) with a parallel groups design (allocation ratio 1:1) using a superiority framework. Participants are randomized to the experimental intervention (85% TiB restriction) or an attention-matched control intervention (maintain habitual TiB). Both groups are balanced on age and amyloid deposition measured with Pittsburgh Compound B (PiB) [29] using positron emission tomography (PET). Participants complete overnight sleep studies, functional neuroimaging scans, morning blood draws, and cognitive testing in the laboratory at two time points, before and after the 4-week intervention.

Study setting

ALPS is conducted at the University of Pittsburgh School of Medicine Department of Psychiatry. The ALPS study is overseen by the principal investigator (KAW) with clinical, safety, and methodological oversight from the co-investigators (BJL, HJA, TKK, AMW, MAG, ADC, and DJB) who make key decisions about the conduct and implementation of the study. Data and safety monitoring is led by the data manager (MEF) and unblinded interventionist staff (YCD). Additional oversight is provided by a safety officer, appointed by the sponsor agency, the National Institute on Aging (NIA), as well as the NIA program officer. Participants are recruited primarily from the Greater Pittsburgh area as well as the broader western Pennsylvania area through community advertisements, participant registries, and participant referrals.

Eligibility criteria for participants

Inclusion and exclusion criteria are displayed in Table 1. Participants are adults aged 65–85 with poor sleep, defined as having a mean sleep efficiency ≤ 92% and wake time after sleep onset > 20 min based on 1 week of sleep diary and actigraphy (ActiGraph CentrePoint® Insight Watch). Sleep efficiency for inclusion/exclusion is defined as the proportion of time spent sleeping within the period of time from “lights out” (the time of the intention to fall asleep after getting into bed) to “final awakening” (before getting out of bed), as this definition of sleep efficiency has been shown to distinguish good sleepers from those with insomnia [30], who may benefit most from the TiB restriction intervention. Participants may have insomnia but do not need to have a formal diagnosis of primary insomnia, so long as they meet the sleep efficiency and wake after sleep onset criteria. Additional inclusion criteria for the ALPS study are corrected-to-normal or normal visual and auditory acuity to complete cognitive tasks. Participants are not required to be native English speakers and are eligible so long as they are sufficiently proficient in English to complete the screening and consent process. Participants are asked at the time of screening if they are participating in any other clinical or research interventions, including exercise and sleep interventions, and enrollment is delayed until other interventions are complete. Participants already enrolled in the trial who are prescribed a new medication (i.e., a course of steroids) would be instructed to pause the intervention until the medication course is complete. If there is no anticipated completion date for the new medication, the participant would be withdrawn from the trial.

Table 1.

Exclusion criteria and methods for assessment. * denotes exclusion for participants completing the MRI

	Method
Inclusion criteria
Age 65–85	Interview
Sleep efficiency < 92% and wake time after sleep onset > 20 min	Diary; confirmed with actigraphy
Normal or corrected-to-normal visual and auditory acuity	Interview
Exclusion criteria
Regular work within the hours of 12 am and 6 am	Interview
Near-miss or prior automobile accident “due to sleepiness” within the past 12 months	Interview
Employed as a commercial driver during the study (for example, bus drivers, train engineers, airplane pilots)	Interview
Epworth Sleepiness Scale score > 10	Interview
Telephone Interview for Cognitive Status (TICS) score < 23	Interview
*Contraindications for MRI: metal in the body, claustrophobia, body mass index > 40	Interview
Consumption of >14 alcohol drinks/week or >6 drinks/sitting	Diary
Consumption of >3 caffeine drinks per day	Diary
Apnea/hypopnea index > 15	ApneaLink™ Air
Diagnosis of a central nervous system disease, such as multiple sclerosis, stroke, Parkinson’s disease, Alzheimer’s disease, seizure disorder, delirium or dementia, a loss of consciousness > 24 h, or traumatic brain injury	Medical history and physical
Presence of a chronic condition that significantly affects sleep	Medical history and physical
Current use of medications affecting sleep such as antidepressants, antipsychotic medications, anticonvulsants, and steroids	Medical history and physical
Current use of sedating drugs used at bedtime	Medical history and physical
Current (past 6 months) major depressive disorder, panic disorder, substance use disorders, and alcohol abuse/dependence	Psychiatric interview
Lifetime history of a psychotic disorder or bipolar I disorder	Psychiatric interview

ALPS exclusion criteria are primarily intended to minimize safety risks associated with TiB restriction. Sleep restriction resulting from TiB restriction may exacerbate the conditions and side effects of medications listed in the exclusion criteria. These conditions may also exacerbate the sleepiness-related safety risks of TiB restriction. Accordingly, participants are excluded if they have any sleep disorders other than insomnia based on a clinical interview, have excessive daytime sleepiness (Epworth Sleepiness Scale > 10), or have moderate-severe sleep apnea, defined as an apnea-hypopnea index (AHI) > 15 events per hour based on one night of in-home screening with an ApneaLink™ Air device (ResMed). An AHI of 15 was selected to exclude individuals with moderate to severe sleep disordered breathing, as sleepiness-related safety risks may be greater in those with moderate to severe sleep apnea [31]. Similar to other protocols where sleepiness is a risk [32], participants are excluded if they had a near-miss or prior automobile accident due to sleepiness in the past 12 months or are employed as a commercial driver. Participants are excluded if they have a chronic unstable medical condition that significantly affects sleep or are taking medications that significantly affect sleep, including antidepressants, antipsychotic medications, anticonvulsants, and steroids. Participants may not have any major neurological conditions. They may have mild cognitive impairment based on neuropsychological testing and formal adjudication, but decision-making abilities must be intact based on a score ≥ 23 on the Telephone Interview for Cognitive Status (TICS). Participants may continue taking all non-exclusionary medications, including cognitive enhancers, if they plan to be on a stable dose throughout the study. High consumption of alcohol or caffeine is exclusionary to minimize effects of these substances on sleep and cognition. However, participants meeting alcohol or caffeine exclusion may opt to reduce their caffeine and alcohol intake for at least 1 week prior to participation and throughout the study, confirmed with diary. Participants completing the MRI portion must meet MRI-specific eligibility criteria.

Eligibility criteria for individuals administering the intervention

Interventionists for ALPS have bachelor’s level training and have completed research training modules as required by the University of Pittsburgh through the Collaborative Institutional Training Initiative (CITI) Program. All members of the research team, including the interventionists, receive ongoing research training in avoiding coercion or any undue influence on potential participants. Interventionists and other staff are instructed not to overstate anticipated benefits at any point before or during participant involvement in the study.

Participant timeline

The study flow chart is displayed in Fig. 1. Participants begin with in-person screening (visit 1). If they are deemed eligible based on their in-person screening visit, diary, actigraphy, and apnea assessment, they are enrolled in the study. Enrolled participants complete the PET scan and neuropsychological cognitive adjudication battery at visit 2 (Table 2). They then complete the first pre-intervention visit (visit 3), are randomized to TiB restriction or control, and then follow their assigned TiB schedule for 4 weeks. Participants return for their post-intervention visit (visit 4) and are informed of the study hypotheses during the study debrief and provided with information about how to improve sleep in older adulthood.

Fig. 1.

ALPS study flow chart

Table 2.

Neuropsychological testing battery for cognitive status adjudication

Instrument	Cognitive domain
Hopkins Verbal Learning Test [33] -Revised Form 3	Word list learning
Symbol Digit Modalities Test [34]	Processing speed
Stroop Color and Word Test Adult Version [35]	Inhibition
Benton Judgment of Line Orientation Form V [36]	Visuospatial skills
Rey-Osterrieth Complex Figure Test [37]	Visual memory
Wide Range Achievement Test 5th Edition Word Reading (WRAT) [38]	Reading/pronunciation
Boston Naming Test 2nd Edition [39]	Language/object naming
Everyday Cognition Scale-2 [40]	Subjective cognitive decline
Wechsler Memory Scale IV Logical Memory 1 & 2 [41]	Narrative verbal memory
Digit Span [42]	Working memory
Controlled Oral Word Association Test Animal and Categorical Fluency [43]	Verbal fluency
Halstead-Reitan Trail Making Test Parts A & B [44]	Processing speed and switching
Performance Assessment of Self-Care Skills (PASS-Clinic) Version 4.1 C14-Medication Management & C8-Shopping [45]	Independent daily living tasks

Interventions

The two intervention arms are TiB restriction and an attention-matched control. For both interventions, the participant’s habitual in- and out-of-bed times and TiB are first determined based on the baseline sleep diary. Habitual TiB for the intervention includes the minutes between time into-bed and time out-of-bed as well as all other sleep opportunities over a 24-h period including naps averaged over 1 week.

The habitual sleep schedule and assigned sleep schedule for each group is calculated using formulas (specified below) in Excel, developed by the investigators (KAW, YUD, and DJB). Following calculation of the sleep schedule, interventionists present the sleep schedules to participants using standardized scripts (Appendices A-1 and A-2) at the conclusion of the first overnight visit.

Daily mid-day check-in calls during the first week are made to participants in both groups to monitor and encourage adherence and discuss any barriers to adherence. Both groups receive the same number of check-in calls and complete the same monitoring assessments to ensure that both groups receive the same level of attention and contact with staff. Check-in calls are reduced to once weekly for the subsequent 3 weeks. Check-in calls include questions about in- and out-of-bed times, frequency and duration of naps, episodes of unintended sleep or dozing, alertness and sleepiness levels, side effects, and episodes of falling asleep or almost falling asleep while driving. Scripts for check-in calls are presented in Appendices B-1 and B-2. To avoid inadvertently unblinding blinded staff, only unblinded staff make contact with participants during the intervention period.

TiB restriction intervention

Participants randomized to TiB restriction are prescribed a sleep schedule that limits their TiB to 85% of their habitual TiB. For instance, a participant with 9 h of TiB at baseline will follow a sleep schedule of 7.65 h TiB. To reduce TiB by 15%, TiB is truncated equally by 7.5% at the beginning and end of the night, to keep the sleep midpoint consistent.

The following equations are used to determine the prescribed in- and out-of-bed times for the TiB restriction group, such that the TiB at the beginning and end of the night are equally truncated by 7.5%:

$$\text{In bed time}=\text{baseline in bed time}+0.075\ast \text{baseline minutes in bed}$$ 1

$$\text{Out of bed time}=\text{baseline out of bed time}-0.075\ast \text{baseline minutes in bed}$$ 2

Participants are asked not to spend time in bed or have any napping or sleep opportunities outside of their prescribed in- and out-of-bed times. For participants who reported naps during the baseline, 85% of the nap minutes were added to the prescribed TiB, and in- and out-of-bed times were adjusted accordingly. Participants may request adjustments in the specific clock times up to 30 min to ensure feasibility while maintaining 85% TiB. The prescribed TiB is never <6 h to avoid negative side effects of very short sleep [11].

Control intervention

For the control group, habitual in- and out-of-bed times are calculated and presented to the participant. Participants are asked to maintain their habitual TiB schedule for the 4-week intervention period. This control comparison was selected to compare the active intervention with making no change to one’s sleep schedule. Although the active TiB intervention also consolidates the sleep period to nighttime, we opted not to discourage naps or any daytime sleep opportunities in the control group to avoid inadvertently restricting TiB in the control group as well.

Criteria for discontinuing or modifying allocated interventions

Criteria for discontinuing or modifying the intervention are based on electronic momentary assessment of the Karolinska Sleepiness Scale (KSS) and telephone check-in calls (Table 3). Throughout the intervention, participants are asked each evening to rate how sleepy they were over the course of that day using the electronic Karolinska Sleepiness Scale. If a participant rates their sleepiness as ≥8 for more than 2 consecutive days, the intervention is modified to 90% TiB, and participants are given adjusted in-and-out of bedtimes. If sleepiness ratings ≥ 8 continue, the intervention is modified to 95% TiB or discontinued. Participants are instructed that, in cases of excessive daytime sleepiness, they may take a brief nap (~20 min), preferably before 4 PM, to alleviate symptoms.

Table 3.

Frequency of check-in calls (orange cells) and electronic Karolinska Sleepiness Scale notifications (green cells)

Participants are asked to spontaneously report any sleepiness-related safety issues and are prompted to answer sleepiness-related safety questions during check-in calls that occur daily in the first week of the intervention and weekly thereafter. If the participant gives a positive response to “have you fallen asleep, even briefly, while driving,” the interventionist tells the participant to immediately stop the intervention and goes through each of the steps listed in Table 4. The safety officer is notified within 3 days of the positive response to this question. The safety officer is given details about the participant’s group assignment, sleepiness ratings, and the context in which the incident occurred. The safety officer advises as to whether the participant should continue in the intervention or not and provides wording for sleepiness-related safety precautions to present to the participant. If any other safety-related adverse event occurs, the safety officer is notified. The safety officer determines whether the participant should discontinue the intervention and/or participation in the study.

Table 4.

Safety protocol for a positive response to “Have you fallen asleep even briefly, while driving?”

During check-in call	Ask the participant to elaborate and write down their responses: • Can you tell me more about that? • When was this? • What was going on at the time?
	Administer FIBSER questionnaire for severity of side effects
	Tell participant to immediately stop the intervention until further notice if they are in the intervention group
	Recommend taking a nap to mitigate sleepiness
	Ask if the participant has any questions or concerns
After check-in call	Immediately notify the principal investigator of the positive response to falling asleep while driving and provide: • Details about the participant’s report of the incident • Intervention arm assignment • All diary/and sleepiness data collected before and during intervention
	Notify the safety officer within 3 days. The safety officer will: • Decide whether the participant can continue in the study or not, and if so, under what conditions • Make a recommendation about the study moving forward
	Contact participant to communicate recommendations as worded by the safety officer including: • Clinical referral information • Recommended precautions

In line with an intent-to-treat approach, participants will complete all assessments, regardless of adherence, deviation, or discontinuation from the protocol or intervention. For all participants who discontinue the intervention, the safety officer makes a recommendation as to whether the participant should continue any remaining assessments. Participants who withdraw or are withdrawn from the study will not complete any remaining study procedures or assessments. Data collected prior to the withdrawal will be used in intent-to-treat analyses.

Adherence and retention

Adherence to the intervention is encouraged during check-in calls (daily in the first week and weekly in subsequent weeks). Electronic text reminders to complete diaries and sleepiness ratings and wear the actigraph are sent throughout the intervention period. Barriers to following the TiB schedule or completing research activities are discussed at check-in calls. Retention is encouraged through these check-in calls and through reminder calls prior to each visit.

Recruitment

The main resource for community recruitment is the University of Pittsburgh Clinical and Translational Science Institute participant registry (Pitt+Me). Participants are additionally recruited through the University of Pittsburgh Claude D. Pepper Registry, community advertisements, postcards, community talks, word-of-mouth, and social media advertising.

Screening and consent

Participants deemed eligible based on phone screening are scheduled for an in-person consent screening visit (visit 1). Visit 1 includes a medical history and physical conducted by a licensed physician (i.e., A.M.I.) and a psychiatric interview conducted by a mental health nurse. Participants are consented by a licensed physician. Participants who opt not to complete the PET scan may be consented by an experienced study coordinator trained in providing consent. In addition to the consent, participants are given the sleep diary to complete and the actigraph watch to wear for 1 week. They are given the ApneaLink™ Air device to use for one night at home. Participants return these supplies via mail to eligibility determination. Participants who are excluded based on their ApneaLink™ results are provided clinical contact information for evaluation and treatment of sleep disordered breathing.

Randomization

Participants are randomized to the TiB restriction group or the control group with a stratified permuted block design at a ratio of 1:1. Randomization is stratified on age (<75 or ≥75) and PiB status (positive or negative for amyloid deposition) based on a “quick read” of the PiB PET scan (see PET methods below). The allocation sequence was generated by a biostatistician (M.L.W.). Only unblinded staff (the data manager and interventionists) have access to the Excel file containing the allocation sequence. Unblinded staff determine the group assignment using the allocation sequence file. The unblinded interventionist is responsible for delivering the intervention instructions to participants. Assessment of habitual sleep with diary and actigraphy, eligibility, and enrollment is performed exclusively by unblinded staff to avoid blinded staff obtaining information about sleep schedules. Blinded staff have no access to the allocation sequence and are not included in communications regarding the allocation sequence or other sleep data.

PiB PET scan

For the PiB scan quick read, the four 5-min acquisition frames of [¹¹C]PiB PET images are summed into a single image corresponding to 50–70 min post-injection. The summed images are smoothed with a 5-mm Gaussian kernel and presented to an expert rater (B.J.L.) for visual interpretation of PiB status [46] using PMOD software (v.4.5, Bruker Corp., Billerica, MA, USA). A four-point rating scale is used that permits equivocation (0 = unequivocally negative, 1 = probably negative, 2 = probably positive, 3 = unequivocally positive). For randomization, the four-point rating scale is collapsed into a binary scale of either negative (0–1) or positive (2–3). Participants who do not complete the PET scan are randomized as “PiB status unknown.”

Blinding and allocation concealment

Study staff responsible for data collection of primary and secondary measures pre- and post-intervention are blinded to the intervention assignment. Blinded staff do not participate in data safety monitoring to avoid revealing intervention group assignment. For data and safety monitoring purposes, interventionists, the data manager, and the principal investigator are unblinded to the intervention.

Several measures are taken to avoid unblinding. Sleep diaries and actigraphy are processed by unblinded staff who are responsible for determining the prescribed sleep schedule for each participant. Data containing sleep diary and actigraphy information are stored on separate drives to avoid unblinding. Unblinded study data collection takes place in a separate office space from blinded study staff procedures. Only unblinded staff are included on communications regarding diary, actigraphy, and in-and out-of-bed times. Participants are informed of which staff should remain blinded to the intervention and are asked not to discuss their in- and out-of-bed times with the blinded staff. Similarly, other study staff (i.e., sleep technicians) are reminded not to discuss in- and out-of-bed times with blinded staff. Personnel responsible for collecting, reducing, and processing polysomnography and plasma data are blind to study hypotheses.

Outcome measures

Primary and secondary measures are displayed in Table 5. The assessment grid for when measures are assessed is displayed in Table 6.

Table 5.

Primary and secondary measures

	Primary measures	Secondary measures
Sleep	• NREM slow-oscillation activity (0.5–1 Hz) absolute power	• NREM slow-oscillation activity (0.5–1 Hz) relative power • Delta activity spectral power (1–4 Hz) • Sleep efficiency (PSG, actigraphy, and diary)
fMRI	• Hippocampal activation during memory encoding	• Medial prefrontal-hippocampal connectivity • Medial temporal-hippocampal connectivity • Medial prefrontal activation during memory encoding
Plasma	• Plasma Aβ1–42	• Plasma Aβ composite score
Cognitive performance	• Overnight A-B word pair memory retention	• PACC score • Accuracy and response time on computerized cognitive tasks

Table 6.

Assessment grid. TX, treatment intervention; Wk, week

	Screen	Baseline	Pre-TX	Intervention				Post-TX
	Visit 1	Visit 2	Visit 3	Wk1	Wk2	Wk3	Wk4	Visit 4
Informed consent	X
In-person screen	X
PiB PET scan		X
Adjudication		X
Randomization			X
TiB intervention				X	X	X	X	X
Diary/actigraphy	X			X	X	X	X
Daily check-ins				X
Weekly check-ins					X	X	X
Overnight PSG			X					X
Cognition			X					X
fMRI			X					X
ApoE			X
Plasma Aβ			X					X
Debrief								X

Primary measures

The specific aims for ALPS examine the effect of TiB restriction on four distinct co-primary measures to assess the overall conceptual model: (1) NREM slow-oscillation activity (0.5–1 Hz) absolute power; (2) hippocampal activation during memory encoding; (3) plasma Aβ_1–42; and (4) overnight memory retention. Analysis of the effects of the intervention on each of these primary measures allows for an assessment of the overall conceptual model that an increase in SWA would reduce hippocampal hyperactivation to reduce Alzheimer’s pathology and improve memory performance.

Secondary measures

Secondary measures for each aim are displayed in Table 5, including NREM slow-oscillation activity (0.5–1 Hz) relative power, delta activity (1–4 Hz), sleep efficiency measured through PSG, actigraphy, and diary for specific aim 1. Secondary fMRI measures are medial prefrontal-hippocampal connectivity, medial temporal lobe-hippocampal connectivity, and medial prefrontal activation during memory encoding for specific aim 2. The secondary plasma measure for specific aim 3 is the plasma Aβ composite score. Secondary cognitive outcomes for specific aim 4 are the Preclinical Alzheimer’s Cognitive Composite (PACC) as well as accuracy and response time on all computerized cognitive tasks.

Sleep

NREM sleep physiology is assessed with in-lab polysomnography with quantitative electroencephalography (EEG). A standard montage is used with F3, F4, C3, C4, O2, and O1 as well as Fz to Cz to assess midline frontal slow-wave activity. The PSG record is scored, and spectral analysis is performed using standard methods [4, 22]. The primary dependent measure to be assessed for sleep (specific aim 1) is NREM slow-oscillation activity (0.5–1 Hz) absolute power.

fMRI

Excitotoxic hyperactivity will be assessed through an in-scanner face-name memory task [47, 48], which is used to assess hippocampal activation during memory encoding. This task has been shown to elicit hippocampal hyperactivation in participants with preclinical Alzheimer’s disease [47, 48]. Participants complete this task during a 30-min fMRI scan in a 3-T Prisma scanner in the morning after their overnight sleep study.

In the first task block, participants view face-name pairs one at a time on the screen. In the second task block, they see the same set of face-name pairs along with a new set of pairs not presented in the first block. The mean novel block > repeated block contrast is used to test for memory encoding-related activation and functional connectivity.

Functional images are processed in Statistical Parametric Mapping (SPM12, https://www.fil.ion.ucl.ac.uk/spm/SPM) with standard preprocessing steps: motion correction, normalization to standard anatomical space, and Gaussian smoothing at 8 mm. To assess task-based activation, MarsBaR will be used to extract the mean activation from region of interest masks (hippocampus and medial prefrontal cortex), in the novel > familiar block contrast for each participant for each time point. The primary dependent measure for fMRI (specific aim 2) is novel > familiar task-based activation within the hippocampus during memory encoding. Task-based functional connectivity will be used as a secondary measure with generalized psychophysiological interaction (gPPI) will use the hippocampus as a seed region.

Plasma and biological specimens

Blood draws are completed the morning after each in-lab sleep study by a certified phlebotomist. Blood samples are collected in ethylenediaminetetraacetic acid (EDTA) tubes and BD P100 tubes and processed in line with published methods for Aβ mass spectrometry [49]. All plasma Aβ samples are stored in a 4 °C refrigerator in a locked room. The plasma sample will generate Aβ_1–40, Aβ_1–42, and amyloid precursor protein (APP)_669–711, and an immuno-precipitation mass spectrometry composite biomarker score [49, 50] is computed as the mean of two z-normalized ratios, Aβ_1–40/Aβ_1–42 and APP_669–711/Aβ_1–42. The primary plasma measure for specific aim 3 is plasma Aβ_1–42. The composite score will be used as a secondary measure.

Cognitive performance

The primary cognitive measure (specific aim 4) is overnight retention of A-B word pairs using a computer-based A-B word pair task [51, 52]. Secondary cognitive measures include an adjusted version of the PACC. We created a repeatable cognitive composite based on the PACC previously described [53] that would avoid task overlap with our adjudication battery. Our PACC is comprised of two memory tasks—long delayed free recall score on the California Verbal Learning Test (CVLT) to assess verbal memory and delayed free recall score on the Brief Visuospatial Memory Test (BVMT) to assess spatial memory—as well as the Delis-Kaplan Executive Function System (D-KEFS) [54] Trail-Making task Condition 4 (switching), and the Digit Symbol Substitution Task [55]. Different test stimuli for memory tasks are used at pre- and post-intervention time points. Additional secondary cognitive measures are computer-based cognitive tasks: a task-switching paradigm [7, 56, 57], the Stroop task [57–59], and the brief Psychomotor Vigilance task (b-PVT) [60].

Baseline only measures

Age, sex, education, PiB status, and ApoE status are collected or assessed once pre-intervention. Age, sex, and education are collected during visit 1 as part of sociodemographic information. PiB status is based on a PiB PET scan using [C-11]PiB using published methods [61] on a Siemens mCT PET/CT scanner (see PET scan). Note that, in compliance with the RDRC 21 CFR part 361, PiB PET results are not disclosed to participants, as PiB is not an FDA-approved drug (see Declarations). Blood samples for ApoE genotyping are collected concurrently with visit 3 plasma Aβ samples and processed using standard methods [62].

Sample size

Target recruitment for the ALPS study is 168 participants to complete in-person screening, 116 participants to enroll in the study, and 102 participants to complete study procedures (n = 51 randomized to TiB restriction; n = 51 randomized to control).

Following Cohen [63] which defines small, medium, and large effect sizes as d = 0.2, 0.5, and 0.8, respectively, the ALPS study has ≥80% power to detect small to medium effects (d > 0.44, r > 0.27) on each primary dependent variable after an estimated 12% attrition, n = 102 (51 TiB, 51 control). This power analysis was based on effect sizes from our preliminary data using a similar TiB restriction intervention [22] and published effect sizes from other sleep interventions, as available [7, 64, 65]. The minimum detectable effect size with 80% power for each aim is within the range of observed effect sizes for SWA, d = 0.5–0.66 [8, 22], functional MRI, d = 0.44–0.8 [66–68], and cognitive function, d = 0.63 [7, 8, 64]. As the plasma Aβ measure had not been evaluated in the context of a sleep intervention, the power analysis for aim 3 was based on effects of sleep deprivation on plasma Aβ, d = 0.68 [69]. For the exploratory mediation analysis, the full sample n = 102 is greater than minimum sample sizes recommended (71–90) to detect medium effects with mediation [70]. An alpha level was set to 0.05 for tests of the primary dependent variables for the specific aims. Secondary analyses will control for family-wise type I error using Holm-Sidak multiple comparisons correction within each domain (sleep, fMRI, plasma, and cognition).

Aims and statistical methods

• Aim 1: Examine whether TiB restriction increases SWA after 4 weeks.

• Aim 2: Examine whether TiB restriction improves fMRI indices of excitotoxic hyperactivation.

• Aim 3: Examine whether TiB restriction improves plasma Aβ levels.

• Aim 4: Examine whether TiB restriction is associated with improved cognition.

Summary statistics will be calculated for demographic characteristics, sleep, fMRI, blood plasma, cognition, and baseline status. Distributions will be inspected for normality to inform appropriate transformations to be applied and check for outliers.

Our primary analytic strategy is to use linear mixed models with a random intercept and time effects. This approach allows for missing at random outcome data and thus enables us to use all available data. The raw value for each dependent variable will be included in the linear mixed model. We will regress repeated dependent variables on time point (pre- and post-intervention), group (TiB restriction or control), and the time point × group interaction. A significant interaction will indicate that the change in dependent variables differs between intervention groups. Follow-up contrasts will confirm whether improvements of dependent variables from baseline to 4-week intervention are significant in the TiB restriction group.

For each specific aim 2–4, follow-up regression analyses will test whether change in SWA (from baseline at 4 weeks) is associated with change in each primary measure within the TiB restriction group only. Analyses will be performed across participants, regardless of PiB status, and stratified by PiB status to assess whether significance differs between those who are PiB positive or PiB negative. We will also examine analysis of covariance models and compare results. All analyses will control for age and sex. ApoE status will be a standard covariate for the plasma Aβ aim 3, and education will be a standard covariate for the cognitive aim 4.

Additionally, exploratory mediation analyses will test whether hippocampal hyperactivation and plasma Aβ_1–42 changes mediate the association between change in slow oscillation and change in memory retention, using the MacArthur method suggested for clinical trials [71].

Interim analysis

ALPS is an early-phase trial using a novel experimental intervention to test a comprehensive conceptual model and therefore is not powered for interim analysis. Based on ongoing review of safety and enrollment data, the safety officer or program officer may decide to terminate the trial based on a change in the risk-benefit ratio or recruitment progress.

Analysis methods to handle protocol non-adherence and missing data

Primary analyses will be based on an intent-to-treat approach, including participants regardless of adherence or missing outcome data. However, we will investigate potential missing data mechanisms, and if data are suspected to be not missing at random, we will perform sensitivity analyses stratified by missingness pattern. If data are missing at random, we will use multiple imputation to retain power for testing hypotheses. Per protocol analyses will also be tested including only participants whose diary and actigraphy suggest >70% adherence, whereby intervention participants’ TiB is ≤85% of their habitual TiB on 20 out of 28 nights and control participants’ TiB is >95% of their habitual TiB on 20 out of 28 nights. Sensitivity analyses will additionally control for race/ethnicity, apnea-hypopnea index, cognitive status, insomnia status, and the Cumulative Illness Rating Scale-Geriatric (CIRS-G) comorbidity scale from the medical history and physical, as appropriate.

Data management

Data is collected through paper forms, wearable devices, and web-based sites (WebDataXpress and university licensed Qualtrics). Data are stored on a secured Microsoft SQL Server using Microsoft Access as a user-friendly front end. MRI and PET data are stored on university servers. De-identified actigraphy data is also stored on a password-protected account within the ActiGraph CentrePoint website.

Data monitoring

Data monitoring is conducted by an internal data monitoring team in conjunction with an external safety officer. The internal data monitoring team, consisting of the data manager and unblinded staff, monitors data collection, protocol adherence, safety issues, and data quality at weekly meetings. The data manager reviews all data stored in the relational database to assess for quality on an ongoing basis. The data manager monitors reminders sent to participants to complete remote assessments and ensures that all remote safety and intervention data are collected as expected.

A licensed practicing physician with expertise in sleep medicine and clinical trials was appointed as the safety officer for ALPS as required by the study sponsor (NIA). The safety officer was nominated by the principal investigator. The safety officer completed a conflict-of-interest disclosure in accordance with the funding institute’s requirements and was formally approved by the NIA program officer.

The safety officer is responsible for monitoring overall study progress, safety of participants, and reviewing adverse events. The data and safety monitoring plan specifies that the principal investigator, data manager, and unblinded staff will prepare and submit a data and safety monitoring report to the safety officer on a twice-yearly basis. This includes recruitment progress, data collected on a modified version of the “frequency, intensity, and burden of side effects” questionnaire (FIBSER), data from the Karolinska Sleepiness Scale, and “yes/no” response to a question about sleepiness while driving. Based on enrollment progress and clinical data, the safety officer makes a recommendation as to whether the study should continue as approved and may provide input for any challenges related to enrollment or monitoring. In addition, enrollment data are uploaded in the Clinical Research Operations & Management System (CROMS), as required by the study sponsor (NIA) for the program official to monitor enrollment and discuss adjustments if needed. If enrollment falls below 80% of the projected cumulative enrollment, additional recruitment efforts are initiated, including targeted advertising via social media and additional community talks.

In the event of an unexpected or serious adverse event, the principal investigator will notify the safety officer of the adverse event within 3 days. This includes a “yes” response to the question “have you fallen asleep, or almost fallen asleep, even briefly, while driving in the past week” that will be asked at each phone contact with participants during the intervention. The safety officer will assess whether the adverse event is indeed unexpected or serious and will make a recommendation as to whether the participant should continue in the study and whether the study should continue as approved.

Harms

To minimize risks and monitor safety related to sleepiness and other side effects of the intervention, three steps are implemented for both groups: (1) exclusion of individuals for which sleepiness could be dangerous or worsen an existing medical condition; (2) frequent check-in calls with safety questions; and (3) daily electronic Karolinska Sleepiness Scale ratings. At each phone check-in, participants provide their sleepiness rating to the interventionist, and the interventionist reminds all participants, regardless of sleepiness rating, to avoid driving if they feel sleepy.

At each check-in call, participants are asked to report any side effects that may have arisen from the intervention, including excessive sleepiness and sleepiness while driving. If participants endorse any side effects, a modified version of the FIBSER scale [72] is administered to assess the frequency and severity of side effects due to the intervention. In addition to prompts to report side effects during check-in calls and at the debriefing, participants are asked to spontaneously report any side effects they may have experienced during or after completing the study. Any adverse events are reported to the IRB and the safety officer.

Participants may choose to discontinue the intervention for any reason. Participants may be told to discontinue the intervention at the advice of the principal investigator or safety officer if it is deemed at any time unsafe for the participant to continue the intervention. Participants who discontinue the intervention may complete all assessments, if advisable by the safety officer.

Provisions for post-trial care

As required by the University of Pittsburgh IRB, the ALPS consent form states that if participants have suffered an injury or any harm as a result of participation in the study, they should immediately contact the principal investigator, and that their insurance may be billed for the costs of any emergency treatment. If the participant’s research-related injury requires medical care beyond this emergency treatment, the participant would be responsible for the costs of this follow-up care. There is no plan for additional compensation at this time (see model consent form).

In the case of incidental findings on the participant’s MRI, sleep disorders or sleepiness while driving, incidental findings are reviewed by a practicing physician associated with the University of Pittsburgh Magnetic Resonance Imaging Center, the safety officer, or co-investigator with the relevant expertise and are provided clinical contact information for evaluation and treatment.

Auditing

Formal auditing procedures will not be conducted. Oversight of study conduct and data integrity will be maintained through ongoing monitoring by the internal data monitoring team and periodic independent review by the safety officer and NIA program officer.

Ethics and dissemination

Confidentiality

Study data are kept strictly confidential, and participants’ identities are not revealed in any publication. All participants have participant identification numbers that are used on forms and for data storage purposes. A code linking participants’ personal identifiers to identification numbers is maintained in a separate, secure location in a password-protected file only accessible to study coordinators who require access. All information will be kept in locked files and secure computer databases.

Dissemination policy

A summary of the results of the specific aims will be submitted to ClinicalTrials.gov no later than 1 year after the clinical trial’s completion date. Authorship for peer-reviewed publications will be determined by the principal investigator in accordance with peer-reviewed journal requirements. There is no intended use of professional writers.

Discussion

Sleep and Alzheimer’s pathology are bidirectionally linked. Sleep circuits are vulnerable to Alzheimer’s disease pathology [73, 74], and sleep may promote processes involved in staving off Alzheimer’s disease through processes such as restoring synaptic homeostasis [27], in turn mitigating excitotoxicity [47, 75–77]. The proposed intervention approach is critical to extend beyond the largely cross-sectional literature on sleep and Alzheimer’s pathology [78–80] to determine if cognition and Alzheimer’s disease risk indicators can be improved through sleep. Although it is counterintuitive to restrict TiB to improve cognition and brain health, slow-wave activity and sleep efficiency may be more beneficial to cognition in older adults than increasing sleep duration [22]. Thus, the results of the proposed trial have the potential to inform sleep health recommendations for older populations.

We view the current intervention as being useful and simple to implement for research investigating effects of improved sleep efficiency and slow-wave activity on cognitive and other health outcomes in older adults. The current experimental intervention is delivered by bachelor’s level personnel without formal specialized sleep medicine training. Accordingly, this TiB restriction intervention may have greater feasibility for assessing cognitive and health trajectories in large-scale community-based studies. This intervention may additionally serve as an initial probe to examine whether improved slow-wave activity is associated with a favorable trajectory for cognitive aging and other aspects of health. Additionally, due to its simplicity, this intervention has the potential to be more easily translated to recommendations provided in a clinical setting, if determined to be beneficial.

Limitations and future directions

Although the ALPS TiB restriction intervention is relatively simple for both the interventionist and the participant to implement, an interventionist with less formal training may be less likely to recognize signs of more serious sleep problems. The intervention does not use other techniques often used in clinical settings, such as motivational interviewing, anticipation of challenges, identifying contributors to long TiB, or other more precise customizations of the intervention to the individual.

Extended TiB restriction may have negative consequences on cognitive measures such as psychomotor vigilance by increasing sleepiness. The current study will characterize these sleepiness-related side effects, determine their role in cognitive changes with the intervention, and determine whether sleepiness is attenuated as participants acclimate to the intervention.

As with all behavioral manipulations, it may be difficult to implement the changes outlined in this intervention, particularly in adults with cognitive impairment. Secondary analyses stratified by cognitive status will begin to assess the suitability of this intervention to populations who already exhibit cognitive impairments.

Future larger-scale studies will examine whether maintaining a TiB restriction schedule, similar to the one used in ALPS, reduces risk for cognitive decline over months to years. These studies are essential to fully test whether behaviorally altering sleep architecture has long-term benefits to cognition and brain health.

Trial status

This trial is ongoing as of December 2025. Recruitment began January 2022 and is expected to continue through May 2026. This protocol is based on IRB modification update #24 dated December 16, 2025.

Authors’ contributions

KAW designed the study and obtained funding. BJL, HJA, ADC, TKK, AMW, MAG, MLW, and DJB are co-investigators and are involved in the design and oversight of the imaging, biomarker, biostatistical, inclusion/exclusion, and/or neuropsychological aspects of the study design. KAW, DJB, YD, and RKH designed and finalized the intervention protocol. AMW and KAW designed the neuropsychological testing battery. MEF is responsible for data management and randomization and provided critical writing edits. AMI, MAG, and DJB perform clinical oversight and provided critical writing edits. YD, KD, RFH, and KAW drafted the manuscript. All authors read and approved the final manuscript.

Funding

This research is supported by the National Institute of Health/National Institute on Aging (grant numbers R01 AG068001, K01 AG049879, P01 AG025204) and the University of Pittsburgh Clinical and Translational Science Institute (CTSI) (UL1 TR001857).

Data availability

The final ALPS dataset will be submitted for permanent archive in an online repository. A summary of the results of the specific aims will be submitted to ClinicalTrials.gov within 1 year after the clinical trial’s completion date. All study data will be de-identified.

Declarations

Ethics approval and consent to participate

The study has been approved by the University of Pittsburgh Institutional Review Board (last renewed on December 16, 2025) and is registered with ClinicalTrials.gov (NCT05138848). The principal investigator submits annual reviews to the IRB for continuing approval. All protocol amendments are submitted to the University of Pittsburgh IRB and approved by the safety officer, as appropriate, prior to implementation. The PiB PET portion of this study was approved under the University of Pittsburgh Radioactive Drug Research Committee (RDRC) mechanism.

Competing interests

Dr. Gebara has received research support from Janssen. Dr. Gebara and Dr. Wilckens have received compensation from Otsuka for developing and presenting a disease-state webinar.

Over the past 3 years, Dr. Buysse (DJB) has served as a paid consultant to Sleep Number (not greater than $5000 per year). Consulting has focused on insomnia, measurement of sleep characteristics, and relationships between sleep and health outcomes. DJB is an author of questionnaires including the Pittsburgh Sleep Quality Index, Pittsburgh Sleep Quality Index Addendum for PTSD (PSQI-A), Brief Pittsburgh Sleep Quality Index (B-PSQI), Daytime Insomnia Symptoms Scale, Pittsburgh Sleep Diary, Insomnia Symptom Questionnaire, and RU_SATED (copyrights held by University of Pittsburgh). These instruments have been licensed to commercial entities for fees by the University of Pittsburgh. DJB receives a portion of the licensing fees, paid to him by the University of Pittsburgh. He is also co-author of the Consensus Sleep Diary (copyright held by Ryerson University), which is licensed to commercial entities for a fee by Ryerson University. DJB receives a portion of the licensing fees from the University of Pittsburgh through its agreement with Ryerson University. DJB has received grant/contract support from NIH, PCORI, AHRQ, VA, and Sleep Number for research relating to sleep, insomnia, sleep interventions, and questionnaire development.

Dr. Karikari (TKK) has consulted for Quanterix Corporation, SpearBio Inc., Neurogen Biomarking LLC., and Alzheon and has served on advisory boards for Siemens Healthineers and Neurogen Biomarking LLC., outside the submitted work. He has received in-kind research support from Janssen Research Laboratories, SpearBio Inc., and Alamar Biosciences, as well as meeting travel support from the Alzheimer’s Association and Neurogen Biomarking LLC., outside the submitted work. TKK has received royalties from Bioventix for the transfer of specific antibodies and assays to third party organizations. He has received honoraria for speaker/grant review engagements from the NIH, UPENN, UW‐Madison, the Cherry Blossom symposium, the HABS-HD/ADNI4 Health Enhancement Scientific Program, Advent Health Translational Research Institute, Brain Health conference, Barcelona‐Pittsburgh conference, the International Neuropsychological Society, the Icahn School of Medicine at Mount Sinai, and the Quebec Center for Drug Discovery, Canada, all outside of the submitted work. TKK is an inventor on several patents and provisional patents regarding biofluid biomarker methods, targets and reagents/compositions, that may generate income for the institution and/or self should they be licensed and/or transferred to another organization. These include WO2020193500A1: Use of a ps396 assay to diagnose tauopathies; US 63/679,361: Methods to Evaluate Early-Stage Pre-Tangle TAU Aggregates and Treatment of Alzheimer’s Disease Patients; US 63/672,952: Method for the Quantification of Plasma Amyloid-Beta Biomarkers in Alzheimer’s Disease; US 63/693,956: Anti-tau Protein Antigen Binding Reagents; and 2450702-2: Detection of oligomeric tau and soluble tau aggregates.