Conference PaperPDF Available

Clinical evaluation of the Embrace smartwatch detection capability of generalized tonic-clonic seizures recorded at the ankles

Conference Paper

Clinical evaluation of the Embrace smartwatch detection capability of generalized tonic-clonic seizures recorded at the ankles

Abstract

Rationale: Embrace is the first FDA approved wrist-worn device combining Accelerometer (ACM) and Electrodermal Activity (EDA) to detect and alert to generalized tonic-clonic seizures (GTCSs). A multi-site clinical evaluation of the machine learning classifier embedded in Embrace validated 94.5% sensitivity of the GTCS detector with a false alarm rate typically lower than 0.2/day (Onorati et al, Epilepsia 2017). When applied on real-life data, Embrace yielded comparable performances in different analyses (Caborni et al, 32nd International Epilepsy Congress 2017; Regalia et al, American Epilepsy Society meeting 2017; Onorati et al, 12th European Congress on Epileptology 2018). The ankle has been reported as more comfortable than the wrist by a significant portion of patients (Fedor et al, 54th Annual Meeting of the Society for Psychophysiological Research, 2014), especially for pediatric patients. The aim of this work is to investigate whether the Embrace detection algorithm, designed for wrist data, shows good generalizability to ankle data gathered in clinical settings. Methods: Data were collected during clinical v-EEG monitoring and consist of 191 recordings taken from 86 patients (81 children, average age: 8 y; 5 adults, average age: 22 y) wearing an Empatica E4 device, provided with EDA and 3 axis ACM sensors, at the ankle. Seizures were labeled by two independent epileptologists. EDA and ACM recordings were analyzed off line with the detection algorithm, previously trained only on wrist data. Sensitivity (Se) was computed as the ratio of GTCSs that triggered an alert. False alarm rate (FAR) was computed as the number of alerts not corresponding to GTCS events, divided by the total recorded hours, normalized by 24 hrs. Results: Overall, 150 days of data were recorded, including 15 generalized GTCSs from 10 patients (9 children). 14 GTCSs were successfully detected (Se=93.3%) by the algorithm. The missed seizure (from a 6 y/o female) exhibited a milder ACM pattern. The cumulative FAR was 0.03, roughly corresponding to 1 false alarm per month. Seventy-three patients (85%) did not experience false alarms; 9 patients (10%) had a FAR between 0 and 1; 2 patients (2.5%) had a FAR between 1 and 2, while 2 patients had a FAR higher than 2. Conclusions: The performance of an ACM and EDA-based classifier trained on wrist data was evaluated on ankle data for the first time. Both the sensitivity and the FAR indicate that the classifier is able to generalize well to motor and autonomic seizures patterns sensed at the lower limbs. The much lower FAR observed on ankle data may be due to the lower occurrence of seizure-like patterns at the ankle. Future work will evaluate the performance also in more challenging real-life ambulatory environments.
Embrace
(Empatica Inc., Cambridge, MA) is the first
FDA approved
device combining 3-axis
accelerometer
(ACM) and
electrodermal activity
(EDA) sensors to
detect and alert to
generalized tonic-clonic seizures
(GTCSs) (
Figure 1
).
Clinical evaluation of the Embrace smart watch detection capability of
generalized tonic-clonic seizures recorded at the ankles
C. Caborni1,G.Regalia1,F.Onorati1, R.W. Picard1,2
For the
first time
,the performance of a
GTCS detection algorithm
trained on
wrist data has been evaluated on
ankle data
.
Both the Se and FAR indicate that the
algorithm generalizes well
to motor and
autonomic seizures patterns sensed at
lower calf
,meeting
performances
endpoints
previously accepted by the
FDA
on wrist EDA and ACM data
[5]
.
By
tuning the classifier’s thresholds
purposely on ankle data, we could achieve
the
same Sensitivity
with
lower FAR
(i.e. 0.03,
1false alert per month
).
This study suggests that the back of the lower calf is aviable site for
long term
monitoring of GTCS in pediatric patients
.
Data collection
Patients were admitted to top hospitals EMU centers for
video-EEG
monitoring.They were asked to wear
an
Embrace
at the
ankle
.
The seizures experienced during the recordings were labeled by
two independent epileptologists
,
following the
ILAE 2016 classification
.
The dataset consists of
150 days
of data from
86 patients:
81 children (41 females), from 2months old to 18 y/o;
5adults (4 females) in the age range of 22-25 y/o;
10 patients
(9 children) experienced
15 GTCSs
seizures.
Data analysis
EDA and ACM recordings were analyzed
off-line
with a
detection algorithm
previously
trained only on
wrist data
from clinical settings.
Sensitivity
(Se) was computed as the ratio of GTCSs detected by the classifier with respect to the total
number of seizures experienced.
False alarm rate
(FAR) was computed as the overall number of detections not corresponding to GTCS
events, divided by the total of recorded hours, normalized by 24 hrs.
1. Empatica Inc
2. MIT Media Lab, Massachusetts Institute of Technology, Cambridge, Massachusetts, U.S.A
In this work we investigate whether the
Embrace detection algorithm
, designed for
wrist data, shows good
generalizability to ankle data
gathered in clinical settings
Aim
Figure 1 Embrace GTC seizure detection and alert system. A wristband embedding ACM and EDA sensors uses machine learning to detect an event and sends an alert to an app, which
generates a call via a cloud-based service to designated caregivers.
[1]
Onorati, Regalia et al., 2017, Multi-center clinical assessment of improved wearable multimodal convulsive
seizure detectors, Epilepsia, 8(11):1870-1879.
[2]
Caborni et al., 2017, Convulsive Seizure detection improved by setting different parameters for active/rest
periods, 32nd International Epilepsy Congress, Barcelona.
[3]
Regalia et al, 2017, Real-time Performance in Outpatient Settings for Embrace Seizure Detection System
with User-Adjustable Sensitivity, American Epilepsy Society meeting, Washington DC..
[4]
Performance of a wrist-worn multimodal seizure detection system for more than a year in real-life settings,
12th European Congress on Epileptology, Wien.
[5]
Lai et al, 2018, Performance assessment of Embrace, the first FDA approved smartwatch for convulsive
seizure detection, Partners Against Mortality in Epilepsy, Alexandria (VA).
References
Results
Amulti-site
clinical
evaluation of the machine learning classifier embedded in Embrace validated
94.5% Sensitivity
(Se) of the GTCS detector
with
False Alarm Rate
(FAR) typically
lower than 0.2/day[1]
.
In
outpatients
settings, Embrace obtained comparable performances:
FAR lower than 1per 24 hrs
with
Se >93%[2,3,4]
.
The
ankle
location has been reported as less distracting and more comfortable than the wrist by some patients, especially some
pediatric patients with
autism
,but to date
no clinical evidence
on the effectiveness of mobile GTCS detection at the ankle has been available.
Rationale
Methods
Conclusions
Contacts
Francesco Onorati
fo@empatica.com
Chiara Caborni
cc@empatica.com
Giulia Regalia
gr@empatica.com
Rosalind Picard
rp@empatica.com
Figure 4
Distribution of the individual FAR experienced by the monitored patients (86).
14 out of 15 GTCSs (Se = 93.3%)
were successfully detected by the automated algorithm
(
Figure 2
).
The only
missed seizure
(from a 6 y/o female) exhibited an irregular and
milder ACM
pattern
(
Figure 3
).
The overall
FAR was 0.18
, roughly corresponding to 1 false alarm every 6 days.
85%
of the patients (73) experienced
no false alarm
(
Figure 4
)
.
10%
of the patients (9) had a FAR between 0 and 1 (
Figure 4
).
Only the 5%
of the patients (5 ) had a FAR higher than 1 (
Figure 4
).
Figure 2
Individual count of seizures for each patient.The bars in green represent the detected seizures, while the
bar in black represent the missed seizure from patient #2.
Figure 3
EDA and 3-axes ACM signals of two seizures are shown. The purple areas indicate the seizures.
Left
: the only missed seizure. The ACM pattern shows a discontinuity of the convulsion and also low values
of intensity.
Right
: a detected seizure. The ACM intensity is higher and there are consistent increases of the
EDA during and after the seizure.
... It has been found that skin becomes a better conductor of electricity for a while after exposure to external stimuli [2,5,23,[80][81][82][83][84][85][86][87][88]. EDA methodology is based on innervation the sweat glands in the skin by sympathetic nerves, which control the synthesis of sweat and its excretion [81,89]. ...
Article
Full-text available
Many people live under stressful conditions which has an adverse effect on their health. Human stress, especially long-term one, can lead to a serious illness. Therefore, monitoring of human stress influence can be very useful. We can monitor stress in strictly controlled laboratory conditions, but it is time-consuming and does not capture reactions, on everyday stressors or in natural environment using wearable sensors, but with limited accuracy. Therefore, we began to analyze the current state of promising wearable stress-meters and the latest advances in the record of related physiological variables. Based on these results, we present the concept of an accurate, reliable and easier to use telemedicine device for long-term monitoring of people in a real life. In our concept, we ratify with two synchronized devices, one on the finger and the second on the chest. The results will be obtained from several physiological variables including electrodermal activity, heart rate and respiration, body temperature, blood pressure and others. All these variables will be measured using a coherent multi-sensors device. Our goal is to show possibilities and trends towards the production of new telemedicine equipment and thus, opening the door to a widespread application of human stress-meters.
... Such devices are ideal for psychological research, lie and stress detectors, etc. For daily life, more suitable EDA meters are in the form of watches, wristbands [115][116][117][118], or even smart eyewear [119] and bras [120], which are more practical and comfortable. The imperfection is in decreased sensitivity, while the nervous reactions and change in skin conductivity are more significant in the palm areas, where the stratum lucidum and potential barrier are presented [121][122][123]. ...
Article
Full-text available
Modern Holter devices are very trendy tools used in medicine, research, or sport. They monitor a variety of human physiological or pathophysiological signals. Nowadays, Holter devices have been developing very fast. New innovative products come to the market every day. They have become smaller, smarter, cheaper, have ultra-low power consumption, do not limit everyday life, and allow comfortable measurements of humans to be accomplished in a familiar and natural environment, without extreme fear from doctors. People can be informed about their health and 24/7 monitoring can sometimes easily detect specific diseases, which are normally passed during routine ambulance operation. However, there is a problem with the reliability, quality, and quantity of the collected data. In normal life, there may be a loss of signal recording, abnormal growth of artifacts, etc. At this point, there is a need for multiple sensors capturing single variables in parallel by different sensing methods to complement these methods and diminish the level of artifacts. We can also sense multiple different signals that are complementary and give us a coherent picture. In this article, we describe actual interesting multi-sensor principles on the grounds of our own long-year experiences and many experiments.
ResearchGate has not been able to resolve any references for this publication.