BHI, 2017
bloomlife.com
Variable-length accelerometer features and electromyography to improve accuracy of fetal kicks detection during pregnancy using a single wearable device Marco Altini, Elisa Rossetti, Michiel Rooijakkers, Julien Penders, Dorien Lanssens, Lars Grieten and Wilfried Gyselaers
BHI, 2017
bloomlife.com
FETAL MOVEMENT
Monitoring fetal movement during pregnancy is the most practical and widespread method to assess fetal wellbeing, one of the most important and complex tasks of modern obstetrics.! ! As birth outcomes are strongly linked to the development of fetal conditions during pregnancy, several techniques have been developed to monitor fetal movement up to date!
| 2
BHI, 2017
bloomlife.com
CURRENT CLINICAL PRACTICE Ultrasound: relies on high frequency sound! waves being used to generate an image of the fetus and! can be used only for a limited amount of time due to! safety concerns. Require hospital stays or trained personnel.! ! Continuous cardiotocography: require cumbersome infrastructure and hospital visits, also involving trained personnel to set up the device and process the produced information.!
| 3
BHI, 2017
bloomlife.com
CURRENT CLINICAL PRACTICE Ultrasound: relies on high frequency sound! waves being used to generate an image of the fetus and! can be used only for a limited amount of time due to! safety concerns. Require hospital stays or trained personnel.! ! Continuous cardiotocography: require cumbersome infrastructure and hospital visits, also involving trained personnel to set up the device and process the produced information.! -> Only sporadic checks in the hospital environment
| 4
BHI, 2017
bloomlife.com
NEW PASSIVE SOLUTIONS Accelerometers: Most studies to date involved one single accelerometer placed on the abdomen and reported rather low sensitivity and specificity.! ! Other researchers added a reference accelerometer with the rationale that by monitoring maternal movement! artifacts using an accelerometer placed outside of the abdominal area, fetal movement should be separable from! maternal movement and therefore detected more accurately. !
| 5
Motion inte
BHI,100 2017
bloomlife.com
NEW PASSIVE SOLUTIONS: REFERENCE ACCELEROMETER 0
0
5
10
15
Motion intensity sensor 3
Timestamp (minutes)
200 150 100 50 0 0
5
10
15
Timestamp (minutes)
Reference accelerometer on the back Motion intensity sensor 6
50 40 30
Fetal movements do not appear on the reference accelerometer, while maternal movements typically do!
20 10 0 0
5
10
Timestamp (minutes)
15
| 6
BHI, 2017
bloomlife.com
NEW PASSIVE SOLUTIONS
-> Promising results, still limited practical applicability
| 7
BHI, 2017
bloomlife.com
SINGLE SENSOR
Performance: consistently lower with respect to multiple sensors and reference accelerometers outside of the abdomen area. Higher false positives (harder to discriminate between maternal movements / artifacts and fetal movements)! ! !
How do we reduce false positives?
| 8
BHI, 2017
bloomlife.com
VARIABLE-LENGTH ACCELEROMETER FEATURES AND PHYSIOLOGICAL DATA The proposed techniques aim at reducing false positives by! providing more contextual information related to maternal! movement while still using a single wearable device to cope! with the absence of a reference accelerometer or a more! obtrusive system.! ! To account for different dynamics in maternal and fetal movement, we computed features over two time windows of 0.5 and 4 seconds. The rationale is that short fetal movements should be averaged out over longer time windows but captured over short ones, while maternal movements should appear over windows of both durations.!
| 9
BHI, 2017
bloomlife.com
VARIABLE-LENGTH ACCELEROMETER FEATURES AND PHYSIOLOGICAL DATA Single sensor, short time window 0.3 label
Motion intensity
nothing kick
0.2
0.1
0.0 0
5
10
15
Timestamp (minutes)
Motion intensity
Single sensor, long time window
0.2
0.1
0.0 0
5
10
Timestamp (minutes)
15
| 10
BHI, 2017
bloomlife.com
VARIABLE-LENGTH ACCELEROMETER FEATURES AND PHYSIOLOGICAL DATA Single sensor, short time window 0.3 label
Motion intensity
nothing kick
0.2
0.1
0.0 0
5
10
15
Timestamp (minutes)
Maternal movements appear on both traces!
Motion intensity
Single sensor, long time window
0.2
0.1
0.0 0
5
10
Timestamp (minutes)
15
| 11
BHI, 2017
bloomlife.com
VARIABLE-LENGTH ACCELEROMETER FEATURES AND PHYSIOLOGICAL DATA Single sensor, short time window 0.3 label
Motion intensity
nothing kick
0.2
0.1
0.0 0
5
10
15
Timestamp (minutes)
Fetal movements appear on the short window trace only!
Motion intensity
Single sensor, long time window
0.2
0.1
0.0 0
5
10
Timestamp (minutes)
15
| 12
Motion int
BHI, 2017
bloomlife.com
0.1
VARIABLE-LENGTH ACCELEROMETER FEATURES AND PHYSIOLOGICAL DATA 0.0
0
5
10
15
Timestamp (minutes)
Single sensor, long time window Single sensor, short time window 0.3 label
Motion Motionintensity intensity
nothing
0.2 0.2
kick
0.1 0.1
0.0 0.0
0 0
5 5
10 10
Timestamp (minutes) Timestamp (minutes)
15 15
Maternal movements are more likely to trigger EMG activity!
Single sensor, EHG data Single sensor, long time window
Motion EHG intensity intensity
300
0.2 200
100 0.1
0 0.0
0 0
5 5
10 10
Timestamp (minutes) Timestamp (minutes)
15 15
| 13
BHI, 2017
bloomlife.com
STUDY DESIGN Twenty-two recordings of about 60 minutes duration were collected from 22 pregnant women at different gestational ages during pregnancy, all from week 30 onwards. ! ! Fetal movements ranged between 0 for inactive babies to 315 for hiccups cases. ! ! Measurements were performed using two devices. A research version of the Bloomlife wearable device, configured to acquire two channels EMG at 4096 Hz and triaxial accelerometer data at 128 Hz and the TMSi system including 6 accelerometers, five placed on the abdomen and one on the back.!
| 14
BHI, 2017
bloomlife.com
FEATURES, CLASSIFIER AND CLASS-IMBALANCE Features: low-complexity time domain features (mean, standard! deviation, interquartile range, correlation between axis, sum,! min, max and magnitude).! ! Classification: Random forests. We set the number of features to select at each iteration to the square root of the total number of features.! ! Class imbalance: small number of kicks with respect to the total available data. The optimal ratio between reference class (kicks) and majority class (non-kicks) was determined by crossvalidating and optimizing for F-score. Our optimal balance included all data from the minority class and one fifth of the majority class data !
| 15
BHI, 2017
bloomlife.com
COMPARISONS AND CROSS-VALIDATION We compared four feature sets associated to the two systems used in this study in order to highlight the impact of the novel methods proposed to improve accuracy of a single wearable device:! 1. TMSi (6 accelerometer system) and variable-length features.! 2. Bloomlife (single wearable sensor) and features computed over a short time window only Bloomlife and features computed over both short and long time windows. ! 3. Bloomlife and features computed over both short and long time windows plus EMG features. ! All models were derived and validated using leave one! participant out cross-validation!
| 16
BHI, 2017
bloomlife.com
RESULTS: SENSITIVITY
Sensitivity
Percentage (%)
75
50
25
0 Multi SL
Single S Single SLSingle SLE
Model
| 17
BHI, 2017
bloomlife.com
RESULTS: SENSITIVITY
Sensitivity
Percentage (%)
75
50
25
0 Multi SL
Single S Single SLSingle SLE
Model
| 18
BHI, 2017
bloomlife.com
RESULTS: SENSITIVITY
Sensitivity
Sensitivity does not change much by introducing variablelength and EMG features as the aim of these features is to reduce false positives. !
Percentage (%)
75
50
25
0 Multi SL
Single S Single SLSingle SLE
Model
| 19
BHI, 2017
bloomlife.com
RESULTS: SENSITIVITY
Sensitivity
Sensitivity does not change much by introducing variablelength and EMG features as the aim of these features is to reduce false positives. !
Percentage (%)
75
50
25
0 Multi SL
Single S Single SLSingle SLE
Model
| 20
BHI, 2017
bloomlife.com
RESULTS: POSITIVE PREDICTIVE VALUE Sensitivity
PPV
75
Percentage (%)
Percentage (%)
75
50
50
25
25
0
0 Multi SL
Single S Single SLSingle SLE
Model
Multi SL
Single S Single SLSingle SLE
Model
| 21
BHI, 2017
bloomlife.com
RESULTS: POSITIVE PREDICTIVE VALUE Sensitivity
PPV
75
Percentage (%)
On the other hand, PPV was 0.75 for the 6 sensors system and increased 50 between 0.65 to 0.75 when including variable-length and EMG features 25
0
Percentage (%)
75
50
25
0 Multi SL
Single S Single SLSingle SLE
Model
Multi SL
Single S Single SLSingle SLE
Model
| 22
BHI, 2017
bloomlife.com
RESULTS: POSITIVE PREDICTIVE VALUE Sensitivity
PPV
75
Percentage (%)
On the other hand, PPV was 0.75 for the 6 sensors system and increased 50 between 0.65 to 0.75 when including variable-length and EMG features 25
0
Percentage (%)
75
50
25
0 Multi SL
Single S Single SLSingle SLE
Model
Multi SL
Single S Single SLSingle SLE
Model
| 23
BHI, 2017
bloomlife.com
RESULTS: POSITIVE PREDICTIVE VALUE Sensitivity
PPV
75
Percentage (%)
On the other hand, PPV was 0.75 for the 6 sensors system and increased 50 between 0.65 to 0.75 when including variable-length and EMG features 25
0
Percentage (%)
75
50
25
0 Multi SL
Single S Single SLSingle SLE
Model
Multi SL
Single S Single SLSingle SLE
Model
| 24
Actual
Actual
BHI, 2017 ●
100
bloomlife.com ●
100
RESULTS: TOTAL NUMBER OF KICKS PER RECORDING ●
●
0
● ● ●● ● ● ● ● ● ●● ● ●
●
●
●
● ● ●
0
0
100
200
0
100
Actual vs Detected kicks, Single S
Actual vs Detected kicks, Single SLE ●
●
300
Actual kicks
●
100
200
●
100
●
●
●
● ● ● ● ● ● ● ● ●● ●●● ● ● ●
0
300
Detected kicks
200
0
200
Detected kicks
300
Actual kicks
300
● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●
●
0 100
200
Detected kicks
300
● ● ●● ● ●● ● ● ● ● ● ●● ●
0
100
200
300
Detected kicks
| 25
Actual
Actual
BHI, 2017 ●
100
bloomlife.com ●
100
RESULTS: TOTAL NUMBER OF KICKS PER RECORDING ●
●
0
● ● ●● ● ● ● ● ● ●● ● ●
●
●
●
● ● ●
0
0
100
200
0
100
Actual vs Detected kicks, Single S
Actual vs Detected kicks, Single SLE ●
●
300
Actual kicks
●
100
200
●
100
●
●
●
● ● ● ● ● ● ● ● ●● ●●● ● ● ●
0
300
Detected kicks
200
0
200
Detected kicks
300
Actual kicks
300
● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●
●
0 100
200
Detected kicks
300
● ● ●● ● ●● ● ● ● ● ● ●● ●
0
100
200
300
Detected kicks
| 26
Actual
Actual
BHI, 2017 ●
100
bloomlife.com ●
100
RESULTS: TOTAL NUMBER OF KICKS PER RECORDING ●
●
0
● ● ●● ● ● ● ● ● ●● ● ●
●
●
●
● ● ●
0
0
100
200
0
100
Actual vs Detected kicks, Single S
Actual vs Detected kicks, Single SLE ●
●
300
Actual kicks
●
100
200
●
100
●
●
●
● ● ● ● ● ● ● ● ●● ●●● ● ● ●
0
300
Detected kicks
200
0
200
Detected kicks
300
Actual kicks
300
● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●
●
0 100
200
Detected kicks
300
● ● ●● ● ●● ● ● ● ● ● ●● ●
0
100
200
300
Detected kicks
| 27
Actual
Actual
BHI, 2017 ●
100
bloomlife.com ●
100
RESULTS: TOTAL NUMBER OF KICKS PER RECORDING ●
●
0
● ● ●● ● ● ● ● ● ●● ● ●
●
●
●
● ● ●
0
0
100
200
0
100
Actual vs Detected kicks, Single S
Actual vs Detected kicks, Single SLE ●
●
300
Actual kicks
●
100
200
●
100
●
●
●
● ● ● ● ● ● ● ● ●● ●●● ● ● ●
0
300
Detected kicks
200
0
200
Detected kicks
300
Actual kicks
300
● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●
●
0 100
200
Detected kicks
300
● ● ●● ● ●● ● ● ● ● ● ●● ●
0
100
200
300
Detected kicks
Less overdetections at the recording level as well
| 28
BHI, 2017
bloomlife.com
CONCLUSIONS We proposed a method to improve the accuracy of fetal kicks detection during pregnancy using a single wearable device placed on the abdomen. ! ! Including variable-length accelerometer features, short fetal movement is averaged out over longer time windows but captured over short ones, while maternal movements of greater intensity appear over windows of both durations. As a result, a single wearable device can be used to better discriminate fetal and maternal movement without the need for a reference accelerometer (11% improvement in PPV).!
| 29
BHI, 2017
bloomlife.com
Thank you
Marco Altini, PhD Head of Data Science |
[email protected]
