Deep Learning to Predict Falls in Older Adults Based on Daily-Life Trunk Accelerometry
Early detection of high fall risk is an essential component of fall prevention in older adults. Wearable sensors can provide valuable insight into daily-life activities; biomechanical features extracted from such inertial data have been shown to be of added value for the assessment of fall risk. Bod...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2018-05-01
|
| Series: | Sensors |
| Subjects: | |
| Online Access: | http://www.mdpi.com/1424-8220/18/5/1654 |
| _version_ | 1811306677939470336 |
|---|---|
| author | Ahmed Nait Aicha Gwenn Englebienne Kimberley S. van Schooten Mirjam Pijnappels Ben Kröse |
| author_facet | Ahmed Nait Aicha Gwenn Englebienne Kimberley S. van Schooten Mirjam Pijnappels Ben Kröse |
| author_sort | Ahmed Nait Aicha |
| collection | DOAJ |
| description | Early detection of high fall risk is an essential component of fall prevention in older adults. Wearable sensors can provide valuable insight into daily-life activities; biomechanical features extracted from such inertial data have been shown to be of added value for the assessment of fall risk. Body-worn sensors such as accelerometers can provide valuable insight into fall risk. Currently, biomechanical features derived from accelerometer data are used for the assessment of fall risk. Here, we studied whether deep learning methods from machine learning are suited to automatically derive features from raw accelerometer data that assess fall risk. We used an existing dataset of 296 older adults. We compared the performance of three deep learning model architectures (convolutional neural network (CNN), long short-term memory (LSTM) and a combination of these two (ConvLSTM)) to each other and to a baseline model with biomechanical features on the same dataset. The results show that the deep learning models in a single-task learning mode are strong in recognition of identity of the subject, but that these models only slightly outperform the baseline method on fall risk assessment. When using multi-task learning, with gender and age as auxiliary tasks, deep learning models perform better. We also found that preprocessing of the data resulted in the best performance (AUC = 0.75). We conclude that deep learning models, and in particular multi-task learning, effectively assess fall risk on the basis of wearable sensor data. |
| first_indexed | 2024-04-13T08:49:47Z |
| format | Article |
| id | doaj.art-51e8cc958165442aa153a2b64d6f59e3 |
| institution | Directory Open Access Journal |
| issn | 1424-8220 |
| language | English |
| last_indexed | 2024-04-13T08:49:47Z |
| publishDate | 2018-05-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Sensors |
| spelling | doaj.art-51e8cc958165442aa153a2b64d6f59e32022-12-22T02:53:31ZengMDPI AGSensors1424-82202018-05-01185165410.3390/s18051654s18051654Deep Learning to Predict Falls in Older Adults Based on Daily-Life Trunk AccelerometryAhmed Nait Aicha0Gwenn Englebienne1Kimberley S. van Schooten2Mirjam Pijnappels3Ben Kröse4Department of Computer Science, Amsterdam University of Applied Sciences, 1091 GM Amsterdam, The NetherlandsHuman Media Interaction, University of Twente, 7522 NH Enschede, The NetherlandsNeuroscience Research Australia, University of New South Wales, Sydney 2031, AustraliaDepartment of Human Movement Sciences, Vrije Universiteit Amsterdam, 1081 HV Amsterdam, The NetherlandsDepartment of Computer Science, Amsterdam University of Applied Sciences, 1091 GM Amsterdam, The NetherlandsEarly detection of high fall risk is an essential component of fall prevention in older adults. Wearable sensors can provide valuable insight into daily-life activities; biomechanical features extracted from such inertial data have been shown to be of added value for the assessment of fall risk. Body-worn sensors such as accelerometers can provide valuable insight into fall risk. Currently, biomechanical features derived from accelerometer data are used for the assessment of fall risk. Here, we studied whether deep learning methods from machine learning are suited to automatically derive features from raw accelerometer data that assess fall risk. We used an existing dataset of 296 older adults. We compared the performance of three deep learning model architectures (convolutional neural network (CNN), long short-term memory (LSTM) and a combination of these two (ConvLSTM)) to each other and to a baseline model with biomechanical features on the same dataset. The results show that the deep learning models in a single-task learning mode are strong in recognition of identity of the subject, but that these models only slightly outperform the baseline method on fall risk assessment. When using multi-task learning, with gender and age as auxiliary tasks, deep learning models perform better. We also found that preprocessing of the data resulted in the best performance (AUC = 0.75). We conclude that deep learning models, and in particular multi-task learning, effectively assess fall risk on the basis of wearable sensor data.http://www.mdpi.com/1424-8220/18/5/1654accidental fallsolder adultsmachine learningneural networksconvolutional neural networklong short-term memoryaccelerometry |
| spellingShingle | Ahmed Nait Aicha Gwenn Englebienne Kimberley S. van Schooten Mirjam Pijnappels Ben Kröse Deep Learning to Predict Falls in Older Adults Based on Daily-Life Trunk Accelerometry Sensors accidental falls older adults machine learning neural networks convolutional neural network long short-term memory accelerometry |
| title | Deep Learning to Predict Falls in Older Adults Based on Daily-Life Trunk Accelerometry |
| title_full | Deep Learning to Predict Falls in Older Adults Based on Daily-Life Trunk Accelerometry |
| title_fullStr | Deep Learning to Predict Falls in Older Adults Based on Daily-Life Trunk Accelerometry |
| title_full_unstemmed | Deep Learning to Predict Falls in Older Adults Based on Daily-Life Trunk Accelerometry |
| title_short | Deep Learning to Predict Falls in Older Adults Based on Daily-Life Trunk Accelerometry |
| title_sort | deep learning to predict falls in older adults based on daily life trunk accelerometry |
| topic | accidental falls older adults machine learning neural networks convolutional neural network long short-term memory accelerometry |
| url | http://www.mdpi.com/1424-8220/18/5/1654 |
| work_keys_str_mv | AT ahmednaitaicha deeplearningtopredictfallsinolderadultsbasedondailylifetrunkaccelerometry AT gwennenglebienne deeplearningtopredictfallsinolderadultsbasedondailylifetrunkaccelerometry AT kimberleysvanschooten deeplearningtopredictfallsinolderadultsbasedondailylifetrunkaccelerometry AT mirjampijnappels deeplearningtopredictfallsinolderadultsbasedondailylifetrunkaccelerometry AT benkrose deeplearningtopredictfallsinolderadultsbasedondailylifetrunkaccelerometry |