Population-Specific Recalibration, Not Sensor Type or Placement, Determines Wearable Accuracy for Walking Energy Expenditure After Stroke

Manufacturer algorithms that estimate walking energy expenditure from wearable sensors show limited accuracy in stroke survivors. This study evaluated whether population-specific recalibration corrects this inaccuracy across three sensors that differ in placement, algorithm and physiological channels. Data recorded during a walking task with indirect calorimetry in adults with post-stroke sequelae were extracted from a health data warehouse. An ActiGraph accelerometer was worn at five placements, an RT6 accelerometer at the hip and a SenseWear Armband on both arms. Each manufacturer algorithm was compared with a recalibrated model built for each sensor by leave-one-participant-out regression on signal and clinical characteristics. An Armband variant added skin temperature, near-body temperature and galvanic skin response. Every manufacturer algorithm performed poorly, with R² ranging from −2.04 for the ActiGraph at the hip to 0.13 for the RT6. Every recalibrated model reached an R² between 0.58 and 0.73, with a mean absolute error of 2.7 to 3.5 kcal. The multisensor Armband variant raised R² from 0.59 to 0.73 at the affected arm but not at the non-affected arm. Population-specific recalibration, rather than the choice of sensor or placement, determined accuracy. Recalibrated wearable sensors can therefore support home monitoring of walking energy expenditure after stroke.

Authors

Publication Details

Journal
Preprints.org
Published
2026-11-18
DOI
https://doi.org/10.20944/preprints202611.1827.v1
Primary Topic
Context-Aware Activity Recognition Systems
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

Population-Specific Recalibration, Not Sensor Type or Placement, Determines Wearable Accuracy for Walking Energy Expenditure After Stroke

Maxence Compagnat, Stéphane Mandigout, Lucas Detto, Alix Poyet et al.
Preprints.org
Context-Aware Activity Recognition Systems
preprint

Population-Specific Recalibration, Not Sensor Type or Placement, Determines Wearable Accuracy for Walking Energy Expenditure After Stroke

Maxence Compagnat, Stéphane Mandigout, Lucas Detto, Alix Poyet, Jean-Christophe Daviet
preprint en

Abstract

Manufacturer algorithms that estimate walking energy expenditure from wearable sensors show limited accuracy in stroke survivors. This study evaluated whether population-specific recalibration corrects this inaccuracy across three sensors that differ in placement, algorithm and physiological channels. Data recorded during a walking task with indirect calorimetry in adults with post-stroke sequelae were extracted from a health data warehouse. An ActiGraph accelerometer was worn at five placements, an RT6 accelerometer at the hip and a SenseWear Armband on both arms. Each manufacturer algorithm was compared with a recalibrated model built for each sensor by leave-one-participant-out regression on signal and clinical characteristics. An Armband variant added skin temperature, near-body temperature and galvanic skin response. Every manufacturer algorithm performed poorly, with R² ranging from −2.04 for the ActiGraph at the hip to 0.13 for the RT6. Every recalibrated model reached an R² between 0.58 and 0.73, with a mean absolute error of 2.7 to 3.5 kcal. The multisensor Armband variant raised R² from 0.59 to 0.73 at the affected arm but not at the non-affected arm. Population-specific recalibration, rather than the choice of sensor or placement, determined accuracy. Recalibrated wearable sensors can therefore support home monitoring of walking energy expenditure after stroke.

Preprints.org
Context-Aware Activity Recognition Systems
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon