Objective: To evaluate changes in the degree of axial postural abnormalities (APAs) using the AutoPosturePD software during standing and walking under single/dual-task conditions.

Background: We recently developed AutoPosturePD, a platform grounded in artificial intelligence’s deep learning technology, to assess the degrees of trunk flexion from RGB-D images and during dynamic conditions, following the MDS Task Force criteria.

Method: Seventeen PD with APAs (Pisa syndrome and camptocormia) were consecutively enrolled at the University Hospital of Verona and Turin, Italy. We collected several clinical and demographical characteristics, and patients were videotaped for 3 minutes patients to quantify the degree of trunk flexion. Four conditions were evaluated: a) quiet standing single task (s.standing); b) standing while reading a text (standing dual task, d.standing); c) walking back and forth without speaking or performing other tasks (walking single task, s.walking); and d) walking back and forth counting back aloud from 100 to 0, subtracting 7 each time (walking dual task, d.walking).

Results: We found a significant difference in all the postural angles during the performance of the d.standing when compared to the s.standing, with worse anterior trunk flexion in d.standing at the thoracic fulcrum(p<0.001), anterior trunk flexion at the lumbar fulcrum(p=0.002), and lateral trunk flexion(p=0.027).We observed a significant posture change within the s.standing and d.standing tasks over time only for anterior trunk flexion at the lumbar fulcrum (p: 0.0048). The curves of lateral trunk flexion showed a slight tendency toward worsening over time, both in the s.standing and d.standing, without showing a statistically significant difference.We found a trend of worsening of the anterior trunk flexion at the lumbar fulcrum during the performance of the d.walking when compared to the s.walking, albeit not reaching the significant statistical threshold(p=0.088). No other postural angles showed significantly different behavior comparing walking in single/dual tasks. No effect of time in changes of postural angles while walking was found during the 3-minute tests.

Conclusion: Three-minute, static stance, dual-task reading condition may worsen APA. This evaluation deserves to be further confirmed in a larger sample of PD patients.

References: [1] Artusi CA, Geroin C, Imbalzano G, Camozzi S, Aldegheri S, Lopiano L, Tinazzi M, Bombieri N. Assessment of Axial Postural Abnormalities in Parkinsonism: Automatic Picture Analysis Software. Mov Disord Clin Pract. 2023 Feb 20;10(4):636-645. doi: 10.1002/mdc3.13692. PMID: 37070056; PMCID: PMC10105105.

[2] Aldegheri S, Artusi CA, Camozzi S, Di Marco R, Geroin C, Imbalzano G, Lopiano L, Tinazzi M, Bombieri N. Camera- and Viewpoint-Agnostic Evaluation of Axial Postural Abnormalities in People with Parkinson’s Disease through Augmented Human Pose Estimation. Sensors (Basel). 2023 Mar 16;23(6):3193. doi: 10.3390/s23063193. PMID: 36991904; PMCID: PMC10058715.

[3] Geroin C, Artusi CA, Nonnekes J, Aquino C, Garg D, Dale ML, Schlosser D, Lai Y, Al-Wardat M, Salari M, Wolke R, Labou VT, Imbalzano G, Camozzi S, Merello M, Bloem BR, Capato T, Djaldetti R, Doherty K, Fasano A, Tibar H, Lopiano L, Margraf NG, Moreau C, Ugawa Y, Bhidayasiri R, Tinazzi M; International Parkinson and Movement Disorders Society Task Force on Postural Abnormalities. Axial Postural Abnormalities in Parkinsonism: Gaps in Predictors, Pathophysiology, and Management. Mov Disord. 2023 May;38(5):732-739. doi: 10.1002/mds.29377. Epub 2023 Apr 20. PMID: 37081741.

« Back to 2024 International Congress

MDS Abstracts - https://www.mdsabstracts.org/abstract/dynamic-video-assessment-of-axial-postural-abnormalities-in-parkinsonism-feasibility-and-pilot-analysis/