The onset of muscle overactivity (MO) is one of the consequences of upper motor neuron lesion (UMNL) [1]. Recent literature highlights the key role of surface electromyography (sEMG) in identifying the presence or the absence of MO [2]. A custom algorithm has been recently developed by our group to automatically identify MO in the absence of any joint motion, i.e. at rest [3]. The aim of the current study was to determine its validity in detecting MO at rest from in-field sEMG measurements in acute patients. Study design: validation study; index test v. gold standard comparison. Data: Over 400 hours of data were acquired in [3] from 14 biceps brachii (BB) muscles of 10 bedridden acute neurological patients with arm plegia. Wearable sEMG probes (Mini Wave Plus, Cometa, Italy) with onboard memory and embedded 3D accelerometer were used. Index test : MO automatic classification (y/n) of 10s-lasting sEMG epochs. The algorithm [3] applies a preliminary set of quality controls both in frequency domain and in time domain to discard corrupted data (e.g. due to electromagnetic interferences, partial electrode detachment, artifacts). Next, the presence of muscle activity is assessed using a wavelet- based enhancement of the matched filter technique [4]. Gold Standard: MO visual classification (y/n) of the same epochs processed by the algorithm, performed by two expert assessors. A two-way table was prepared. Sensitivity (Se), specificity (Sp), positive and negative predictive values (PPV, NPV), and overall diagnostic accuracy were computed along with their 95% confidence intervals. Sample size: According to sample size calculation for diagnostic studies [5], assuming a 25% prevalence of MO [3], expected Se = Sp =80% and a ±2.5% CI width, a total of 4000 epochs should be analysed. Preliminary results on 704 epochs distributed among subjects are presented. Of these, 513 epochs had complete absence of movement (e.g., due to nursing activities), as detected by the accelerometer. The resulting two-way table is presented in Fig. 1a. An example of data is presented in Fig. 1 b. Se was 87% (81-89%), Sp was 86% (81-89%), PPV was 79% (73-85%) and NPV was 91% (87-94%). The overall diagnostic accuracy was 86% (83-89%). These preliminary results support the use of wearable sEMG sensors with the developed algorithm for in-field assessment of MO at rest. False negative classifications were often due to activity staring at the end of the 10 s epoch. The assessment and monitoring of MO in acute and subacute patients with UMNL could improve the quality of care by suggesting early treatments to lessen long-term complications. It is worth highlighting that the identification and quantification of MO does not provide a direct clinical diagnosis in any way. Rather, it allows assessing or monitoring a pathophysiological condition.