Freezing of gait assessment with inertial measurement units and deep learning: effect of tasks, medication states, and stops

Abstract

Background: Freezing of gait (FOG) is an episodic and highly disabling symptom of Parkinson's Disease (PD). Traditionally, FOG assessment relies on time-consuming visual inspection of camera footage. Therefore, previous studies have proposed portable and automated solutions to annotate FOG. However, automated FOG assessment is challenging due to gait variability caused by medication effects and varying FOG-provoking tasks. Moreover, whether automated approaches can differentiate FOG from typical everyday movements, such as volitional stops, remains to be determined. To address these questions, we evaluated an automated FOG assessment model with deep learning (DL) based on inertial measurement units (IMUs). We assessed its performance trained on all standardized FOG-provoking tasks and medication states, as well as on specific tasks and medication states. Furthermore, we examined the effect of adding stopping periods on FOG detection performance. Methods: Twelve PD patients with self-reported FOG (mean age 69.33 +/- 6.28 years) completed a FOG-provoking protocol, including timed-up-and-go and 360-degree turning-in-place tasks in On/Off dopaminergic medication states with/without volitional stopping. IMUs were attached to the pelvis and both sides of the tibia and talus. A multi-stage temporal convolutional network was developed to detect FOG episodes. FOG severity was quantified by the percentage of time frozen (%TF) and the number of freezing episodes (#FOG). The agreement between the model-generated outcomes and the gold standard experts' video annotation was assessed by the intra-class correlation coefficient (ICC). Results: For FOG assessment in trials without stopping, the agreement of our model was strong (ICC(%TF) = 0.92 [0.68, 0.98]; ICC(#FOG) = 0.95 [0.72, 0.99]). Models trained on a specific FOG-provoking task could not generalize to unseen tasks, while models trained on a specific medication state could generalize to unseen states. For assessment in trials with stopping, the model trained on stopping trials made fewer false positives than the model trained without stopping (ICC(%TF) = 0.95 [0.73, 0.99]; ICC(#FOG) = 0.79 [0.46, 0.94]). Conclusion: A DL model trained on IMU signals allows valid FOG assessment in trials with/without stops containing different medication states and FOG-provoking tasks. These results are encouraging and enable future work investigating automated FOG assessment during everyday life.

Automatic Detection and Assessment of Freezing of Gait Manifestations

“…Sample-based methods treat FOG detection as an action segmentation task [9], [29], [30]. These approaches distinguish between FOG and non-FOG on the sample level by generating one output for each input sample.…”

“…Such sample-to-sample prediction eliminates the need for pre-defined window sizes and majority voting, which allows for more fine-grained activity detection [31]. In a recent study, we introduced the multi-stage temporal convolutional network combined with a many-to-one training scheme (MS-TCN) [9]. This temporal convolutional neural network architecture modified the training procedure of the multi-stage temporal convolutional network [32], initially proposed for video action segmentation, to improve FOG detection performance.…”

“…Various methods have been proposed to automatically detect and assess FOG using wearable sensor data obtained through IMUs [9], [13]- [18]. IMUs could record the movement of the associated body segment as a time series of 3-axis acceleration and angular velocity.…”

“…Although described as a single phenomenon, FOG is not univocal and can be expressed as different manifestations, namely: 1) episodic rapid shuffling with very short steps and poor clearance of the feet, 2) trembling in place visible as alternating tremulous oscillations in the legs with minimal or no forward progression, and 3) complete akinesia with minimal or no visible movement in the lower limbs [8]. However, whether or not shuffling should be included in the definition of FOG is being debated given that there is still forward progression of the feet [9]. As the etiology of the different manifestations likely differs and, as such, may respond differently to therapy, developing an objective assessment of the FOG manifestations will improve our understanding of this complex symptom and help guide appropriate treatment [8].…”

“…We adjusted three state-ofthe-art FOG detection algorithms to the FOG manifestations detection task. These algorithms served as a baseline for comparison with our previously validated FOG manifestation detection algorithm that was not specifically trained to detect manifestations [9]. To quantify FOG manifestation severity, we calculated the percentage time frozen (%TF) as per previous work [10], [11] and the percentage time frozen of each manifestation.…”