OSSMM: An Open-Source Sleep Monitor and Modulator

1 st Jonny Giordano, 2 nd Fergal Stapleton, 3 rd Gabriel Palma, 4 th Barak A. Pearlmutter · 2026

Read OSSMM: An Open-Source Sleep Monitor and Modulator by 1 st Jonny Giordano, 2 nd Fergal Stapleton, 3 rd Gabriel Palma, 4 th Barak A. Pearlmutter (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-18588.txt.

Introduction

Jonny Giordano and three co-authors present OSSMM, a 2026 open-source sleep monitor and modulator built as a low-cost wearable headband plus an Android app. The headband needs no gels or disposable electrodes and captures movement, pulse, eye movement, and putative EEG signals wirelessly. A fifteen-night single-participant test supports four-stage sleep classification with macro F1 of 0.770 and accuracy of 0.776 against a validated non-contact monitor. Cheap reusable electrodes from fitness chest straps drive classification, even from just two frontal electrodes with no dedicated ground.

Catalog record

CREATOR1 st Jonny Giordano, 2 nd Fergal Stapleton, 3 rd Gabriel Palma, 4 th Barak A. Pearlmutter
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE33296
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2605.18588

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

OSSMM: An Open-Source Sleep Monitor and Modulator 1 st Jonny Giordano, 2 nd Fergal Stapleton, 3 rd Gabriel Palma, 4 th Barak A. Pearlmutter Abstract We present the Open-Source Sleep Monitor and Modulator (OSSMM), an open-source hardware and software platform for accessible sleep research. The OSSMM comprises a small wearable headband built from 3D prints and affordable commercial-off-the-shelf (COTS) components at a material cost under €40, supported by a companion Android application. The system requires no conductive gels, disposable electrodes, or specialized equipment, and captures multiple biosignals — movement, pulse, electrooculography (EOG), and putative electroencephalography (EEG) — with wireless connectivity for data storage and potential sleep modulation capability via an onboard vibration motor. A proof-of-concept single-participant evaluation across 15 nights demonstrated that the captured biosignals support four-stage sleep classification (Wake, Light Sleep, Deep Sleep, REM) using conventional machine learning methods, with the best-performing model achieving a Macro F1-score of 0.770 and accuracy of 0.776 against a validated non-contact sleep monitor ( κ \kappa =0.63 with PSG). Two technical findings are of particular note. First, inexpensive, reusable conductive thermoplastic polyurethane (CTPU) electrodes from commercial fitness chest straps captured a differential signal whose spectral properties in canonical EEG frequency bands, including signatures consistent with sleep spindles, are the principal features driving classification. Second, this signal is obtained from just two frontal electrodes without a dedicated ground reference, suggesting that practical sleep staging is achievable with simpler configurations than typically employed. All hardware designs, software, and build instructions are openly available to support replication and modification by the research community. I Introduction In 1937, Loomis, Harvey, and Hobart first revealed that sleep is not a homogeneous state, but characterized by periods of distinct neural patterns [ 18 ] . This understanding was extended in 1953 with the identification of Rapid Eye Movement sleep (REM) [ 3 ] . Since then, numerous associations have been uncovered between sleep states and facets of cognition and health [ 34 , 28 , 23 , 29 , 14 ] . The accurate identification of sleep stages is therefore essential for sleep research. Traditionally, sleep staging has relied on manual scoring of polysomnography (PSG) by qualified experts [ 1 ] . While considered tthe gold standard, PSG has significant trade-offs. A full PSG (Type 1) is not ecological, requiring attendance at a dedicated facility, trained staff for set-up and monitoring, while demanding the participant tolerate uncomfortable, restrictive equipment. In the United States, a single in-laboratory PSG costs on average $1,840, while an at-home evaluation (Type 2) costs $1,575 [ 15 ] . Many solutions have been put forth to address these drawbacks: reduced electrode placements, portable or at-home devices, and automatic sleep scoring using machine learning [ 8 , 20 , 31 , 33 , 17 ] . However, the barrier to sleep research remains high. Low-cost, accessible sleep monitoring devices perform poorly compared to PSG , while dedicated devices often cost more than a single night of PSG [ 20 , 25 , 27 ] . Without significant funding, sleep researchers must choose between low-cost methods with poorer accuracies for more participants over longer periods, or high-cost methods with improved accuracies for fewer participants and shorter durations. Concerns also remain for the ease-of-use of take-home equipment, with many systems requiring prior set-up by researchers and the use of special conductive gels or electrode pads. Active sleep monitoring, required for sleep modulation research targeting specific sleep states or neural signatures, is more demanding still. Despite promising results from sleep modulation studies, such as potential therapies for depression and cognitive investigations, progress has been limited [ 34 , 16 , 5 , 35 ] . Consider that nearly 25 years ago one sleep modulation study employed near real-time automatic sleep staging, yet more modern studies returned to manual manipulations of sleep [ 8 , 16 , 26 ] . Therefore, a solution is needed which can provide suitable sleep data collection and accurate sleep stage classification, while being affordable, transparent, and easily modifiable for varied experimental requirements. The Open-Source Sleep Monitor and Modulator (OSSMM) project responds to these needs. The platform’s hardware and software are open-source, and designed for assembly with basic electronics knowledge. The system comprises a wearable electronic headband and a dedicated Android application. The headband collects movement, pulse, electrooculography (EOG), and putative electroencephalography (EEG) data and transmits data wirelessly to the connected smartphone. In this study, we conducted a proof-of-concept technical investigation of the OSSMM platform. We hypothesized that nightly sleep data collected from the OSSMM device would support 4-stage sleep classification, namely for states ”Wake,” ”Light Sleep,” ”Deep Sleep,” and ”REM.” To test this, we employed three conventional machine learning classifiers — Support Vector Machine (SVM), Random Forest (RF), and XGBoost (XGB) — with minimal hyperparameter optimization, applied to features extracted with limited digital signal processing. Sleep stage labels were provided by a reference device, the Somnofy SM-100 by Vitalthings, a validated ( \kappa =0.63 agreement with PSG) non-contact sleep monitor [ 33 ] . This conservative analytical approach was chosen to assess the baseline informational content of the OSSMM signals, rather than to maximize classification performance through model or feature engineering. II Methods II-A Experimental Design A total of 17 nights of sleep data were collected using the OSSMM and Somnofy devices. The recording period was designed to yield a minimum of 14 nights of usable data, with 3 spare to accommodate unforeseen disruptions (e.g. illness, equipment faults). This technical evaluation employed a single-participant design (N=1) with one researcher serving as the sole participant, yielding 15,285 30-second epochs (i.e. observations) across 15 qualifying nights. The single-participant design was chosen as this study constitutes a technical validation of the platform, assessing system functionality and hardware reliability prior to multi-participant validation. This experimental design choice aligns with the IEEE Code of Ethics, avoiding unnecessary risk to participants ahead of full system validation [ 9 ] . Generalizability to a broader population is addressed as a limitation in Section IV. The nightly protocol was as follows: The CTPU (conductive Thermoplastic Polyurethane) electrodes were wiped clean with a damp cloth and dried to remove any oils that may have accumulated from the previous night. The device was donned, powered on, and connected to the companion application. A Google Pixel 9 smartphone running Android 16 ran V1.0.4 of the OSSMM application. The application’s real-time plots were used to verify that movement, pulse, and EOG data were being detected and recorded. A typical bedtime activity (e.g., reading) was then carried out for at least 15 minutes to augment the number of Wake observations. Immediately before attempting sleep, a modified consensus sleep diary was partially completed before sleep [ 4 ] . Upon waking the following morning, data recording was halted, and the headband powered off and removed. The headband was connected to a USB-C charger and the remaining sections of the consensus sleep diary were completed covering nocturnal awakenings, perceived sleep quality, and the headband displacement questions described below. The sleep diary was augmented with two self-report questions: “Did you notice the headband move during your sleep?” and “Did the headband come off during sleep?” These questions were included to inform iterative improvements to the headband design and to monitor

Read the complete text