The sleep disturbances of insomnia include difficulty falling asleep, staying asleep, or inability to fall back to sleep after premature awakenings, or any combination of these, along with daytime sleepiness and dysfunction. Difficulty falling and remaining asleep affected nearly 30% of the US population, with approximately 20% of the general population experiencing occasional difficulties and an additional 10% meeting the clinical criteria for an insomnia disorder [1]. However, this protocol was conducted during the COVID pandemic, when the prevalence of insomnia symptoms increased to 53% and insomnia disorder prevalence jumped from 10% to 17% [2]. Insomnia symptoms have been linked with poorer health across multiple domains and exacts a high societal cost in the workplace, on the health care system, and on an individual’s and family’s quality of life [3-5]. Moreover, side effects from medications used to manage insomnia increase these factors in some cases. Novel approaches that facilitate sleep are critically needed, and some potential neural mechanisms of insomnia create a viable treatment target [6].

The most widely accepted theory of primary insomnia is the hyperarousal hypothesis, which suggests that problems with sleep initiation and maintenance are due largely to the disruptive effects of somatic or cognitive hyperarousal [7,8]. In particular, cortical hyperarousal can emerge from excessive focus on repetitive negative thoughts, including intensive problem-solving, self-reflective rumination, and worry [9-11]. Subjectively, people who struggle with insomnia often make comments such as “I wish I could just turn my mind off” or “I just keep replaying conversations over in my head.” In fact, one of the major features of insomnia is the tendency toward self-reflective rumination and worry [9,12-15]. This internal dialogue contributes to a cycle of self-referential thought and hyperarousal that appears to hinder sleep onset and maintenance [9,16-18]. In fact, worry and sleep disruptions are associated in the general population [3]. In individuals with insomnia, cortical hyperarousal contributes to difficulty initiating and maintaining sleep [19].

Neuroimaging research has shown that internally focused self-reflective processes of this type tend to activate an interconnected system within the brain known as the default mode network (DMN) [20,21]. Moreover, negative ruminative thinking is associated with changes in DMN connectivity and other brain regions associated with cognitive arousal or negative emotion [22]. Patients with insomnia often show abnormalities in the functioning of the DMN that are consistent with the hyperarousal hypothesis [23-25]. The core nodes of the DMN typically include the medial prefrontal cortex, posterior cingulate cortex, precuneus, and bilateral inferior parietal cortex regions but can also include several ancillary smaller regions, including the hippocampus, medial temporal lobes, and other subcortical structures [26]. Individuals with insomnia disorder tend to have increased resting-state functional connectivity (FC) between spatially segregated nodes of the DMN [27,28]. Increased connectivity and activation of the DMN could contribute to the ongoing self-referential processing and internal dialogue that maintains a hyperaroused state and perpetuates difficulties falling and remaining asleep [27,29]. Individuals with insomnia disorder show greater activation of the DMN compared with healthy controls while viewing word lists associated with past, present, and future worries, particularly when the words are self-referential [29].

Mainstream approaches to treating primary insomnia include cognitive behavioral therapy for insomnia and pharmacologic sleep aids, such as hypnotic sedatives [30-33]. While cognitive behavioral therapy for insomnia often helps to reduce ruminative cognitions and is effective at improving sleep for many individuals, there are many who fail to achieve meaningful benefits [34,35]. Similarly, pharmacologic sleep aids also tend to have modest effect sizes [36] and are often associated with unwanted side effects (eg, daytime drowsiness and memory problems) and health-related morbidities [37]. Various noninvasive neuromodulatory approaches have been shown to improve insomnia symptoms, including repetitive transcranial magnetic stimulation (rTMS) [38]. However, low-frequency rTMS has a cortical inhibition effect that has been shown to improve insomnia symptoms, which may be due to hyperarousal across multiple psychological and physiological domains [39].

The proposed protocol aims to bridge the gap between cognitive and neuromodulatory interventions to facilitate sleep onset and longer maintenance by temporarily inhibiting the brain FC that is associated with cortical hyperarousal and presleep ruminative cognitions. Our proposed approach will involve using rTMS prior to bedtime to briefly disrupt the strength of FC among cortical regions of the DMN. As we propose to inhibit connectivity within the DMN, we will use continuous theta burst stimulation (cTBS), which induces long-term depression of cortical neural firing following a sustained stimulation period of 40 seconds [40]. Thus, we hypothesize that the application of cTBS to an easily accessible node of the DMN will suppress the local neural activity of that node and propagate inhibition throughout the DMN, thereby reducing ruminative thinking and worry prior to sleep onset. Decreased DMN connectivity is expected to improve sleep quality and quantity. An increase in slow-wave and rapid eye movement sleep stages has been proposed as mechanisms of rTMS in increasing restorative sleep in individuals with insomnia, so polysomnography (PSG) parameters are the primary outcome measures in the outlined protocol [41]. This phase 1 clinical trial will be the first study to investigate the effects of cTBS targeted to the left inferior parietal node of the DMN on objectively measured sleep outcomes.

This study aims to explore the effects of cTBS on (1) the activation and connectivity of the DMN and (2) sleep outcomes. We hypothesize that modulating the DMN by stimulating a targeted region of the left inferior parietal lobule with a single cTBS administration will decrease FC within the DMN and thereby improve sleep parameters (ie, PSG) relative to an identical sham administration.

This phase 1 clinical trial (ClinicalTrials.gov NCT04953559) will be conducted with a randomized, double-blind, sham-controlled, counterbalanced, crossover design. Participants serve as their own controls as they undergo 2 identical study sessions separated by at least 5 days (ie, washout period). Participants and study personnel who interact with the participants will be blind to the specific condition (active cTBS vs sham) the participant receives during each study visit. Using a prespecified, equally balanced, block randomization procedure, participants will be assigned to the order in which they receive active cTBS and sham stimulation conditions. The randomization list will be generated by a computer random number generator, with the constraint that half of the participants of each biological sex will receive sham first and half will receive the active cTBS first. This design will allow us to examine the intraindividual effects of the active cTBS and establish the superiority of active cTBS compared with sham stimulation in improving PSG sleep outcomes.

The study procedures were completed over 3 separate visits. Screening to determine eligibility and baseline assessments was conducted during visit 1. Eligible participants were asked to wear a wrist actigraph for at least 5 days prior to their first in-laboratory overnight visit (visit 2). Based on a preestablished counterbalanced crossover randomization schedule, participants were then assigned to receive either the active cTBS or sham stimulation for their first overnight visit (visit 2) followed by the alternative intervention during their second overnight visit (visit 3) at least 5 days later. As shown in Figure 1, each in-laboratory session lasted 18 hours, including a 2-hour cognitive testing block, 1 hour of preintervention magnetic resonance imaging (MRI), followed by either active cTBS or sham stimulation, and a postintervention hour of neuroimaging. Participants were then fitted with PSG electrodes and were permitted 8 hours of sleep while undergoing PSG sleep monitoring in the lab. The next morning, participants completed another 2-hour cognitive testing block before being released.

The full study execution takes approximately 38-40 hours with 2-3 hours allocated for baseline screening and enrollment (visit 1) and 18 hours allocated for each overnight visit (visits 2 and 3). Table 1 presents the specific assessment measures administered at each visit.

All study activities and study personnel were approved by the University of Arizona’s Institutional Review Board (IRB) and by the Department of Defense Office of Human Research Oversight (OHRO) in March 2021 (approval 2007900971). Any protocol amendments and reportable new information were disseminated to both IRB and OHRO. All participant information was deidentified for confidentiality. Participants were compensated 500 dollars via check for their participation in the study if they were fully compliant with the study procedures. The study team maintains dissemination control of the final deidentified data set. Final results will only report cumulative population data to respective regulatory and scientific reporting agencies and along with peer-reviewed publications. Access to the deidentified data set will be available upon request to the principal investigator and in conjunction with proper regulatory requirement.

A total of 20 otherwise healthy adults with self-reported sleep disturbances were recruited from the greater Tucson and Phoenix metropolitan areas. Recruitment strategies include flyers posted on community boards, advertisements in local newspapers, sponsored social media posts, and email lists at local universities. Individuals who were recruited from the community were directed to complete a web-based survey in a HIPAA (Health Insurance Portability and Accountability Act)–compliant server. If participants met the initial inclusion criteria, a study member then completed a preliminary phone screening with the participant. Study personnel then scheduled an in-person screening visit for eligible individuals.

During the study design, we conducted a power analysis to determine the sample size necessary to compare mean changes in DMN connectivity before and after cTBS using a within-subjects ANOVA approach. From prior cited work on transcranial magnetic stimulation (TMS) and FC, a total sample of 20 was found to have sufficient power to answer the research questions to include analysis of 2 groups (within-subject active and sham), 2 repeated measurements (before and after TMS), and a correlation between the change in DMN activation and sleep outcome metrics [42,43].

Individuals between the ages of 18-50 years with reported sleep disturbances were included in the study. Individuals were excluded if they reported or exhibited any health conditions beyond insomnia-related symptoms. Exclusionary criteria were based on one or more of the following considerations: the criterion in question is (1) known to alter sleep, (2) known to substantially increase interparticipant variability, (3) known to put the volunteer outside the range of what is considered healthy, or (4) required by regulation. Detailed inclusion and exclusion criteria are referenced in Textbox 1.

We hypothesize that active cTBS will disrupt the within-network and between-network connectivity of the DMN. This hypothesis will be assessed from a 10-minute resting-state FC scan and analyzed using standard procedures in the CONN (v17f or higher). Standard pipelines for preprocessing will be used. Within-network connectivity will be determined by first conducting an independent components analysis of the intrinsic activation patterns across the brain. Twenty independent components will be extracted. Mean activation clusters from each component will be examined for intercorrelation to determine within-network connectivity. Additionally, between-network, region-to-region, and seed-to-voxel analyses will be conducted to compare changes in FC between the 2 stimulation conditions.

With a growing number of people reporting greater sleep disturbances, identifying effective nonpharmacological interventions is an important step in providing more innovative and noninvasive approaches for improving sleep outcomes. While the sample size for this study is small, this is the first study to investigate cTBS effects on the DMN to improve sleep outcomes.

Double-blind randomized controlled trials are the gold standard for scientific inquiry. Therefore, each participant will serve as their own control in a counterbalanced crossover design. Implementing a crossover design in which every participant receives both active and sham cTBS stimulation permits comparison of intraindividual differences in DMN connectivity and sleep quality parameters, as each participant acts as their own control. Because the goal was to examine the effects of cTBS on individuals with sleep problems, no healthy control samples were recruited or necessary to test our hypotheses. Blinding study personnel to the TMS condition also reduces systematic differences in the administration of the stimulation interventions and ostensibly reduces the influence of expectations about treatment effects.

Continuous theta burst was selected as the intervention based on its previously demonstrated inhibitory effects on the cortex [47]. A single 40-second cTBS session inhibits, or suppresses, neural excitability in the cortex for up to 50 minutes or longer [47]. Stimulation of a single node of the DMN is expected to inhibit local activity, with further suppression of cortical excitability propagating throughout the network overall. Inhibition of the left inferior parietal lobule was selected for two primary reasons: (1) prior research showed reduced metabolic activity in this region when individuals were falling asleep and (2) this region facilitates an easy-to-access node of the posterior DMN [48].

The localization of specific brain regions requires prestimulation imaging to account for individual variability in brain structure. For this reason, a critical element of that project was to show that stimulation of a single, easily accessible surface node of the larger DMN could lead to significant alterations in FC within this network. Prior investigations using cTBS have focused primarily on the stimulation of brain regions associated with the prefrontal areas of the DMN. Sleep-related parameters have rarely been collected as outcome measures for cTBS. One prior study examining rTMS effects on the right parietal cortex showed reduced anxiety and insomnia symptoms [49]. However, that study used a stimulation session (of three 1-Hz stimulations every 10 minutes) spread across 10 days. The presently proposed cTBS protocol is aimed at establishing initial research outcomes of the effects of a single burst of cTBS on disrupting network connectivity of the DMN to improve sleep. Thus, only a single stimulation session of both active cTBS and sham is needed to identify the intraindividual effects of cTBS on sleep outcomes.

Targeting a single node of the DMN and conducting only 1 stimulation session are limitations of this protocol to sufficiently answer the research questions and aims proposed regarding DMN and insomnia improvement. Future work could determine the most effective stimulation sites within the DMN and the optimal number of pulses and administrations to optimize this potential intervention for sleep problems. Larger samples are also needed than the small pilot sample collected for this protocol to examine the effects of sex observed in previous studies [38].

Insomnia symptoms linked with many other psychiatric and physiological morbidities have become an increasingly prevalent and important health care and societal concern. Noninvasive neuromodulation can be inexpensive and is a technique accessible to many types of treatment providers. The protocol presented here unifies the prominent aspects of insomnia—physiological and psychological hyperarousal—by targeting a unique neural network, known to be associated with proposed mechanisms of insomnia [8,50].