Borelli@Saints-Pères : D. Keriven Serpollet & R. Barrois Muller

Dimitri Keriven Serpollet and Rémi Barrois Muller present their work at the Borelli@Saints-Pères seminar.

Dimitri Keriven Serpollet

Title: Understanding pilot’s sensorimotricity: from vestibular perception to mental workload assessment

Abstract:  The aeronautical sector is looking for innovations to improve flight safety. In this context, my work focused on two main points. On the one hand, we analyzed the differences in perception between different planes of space in the absence of visual cues, keeping in mind the issue of sensory illusions. We showed that a cohort of professional helicopter pilots were more accurate in the frontal plane than in the sagittal plane (p<0.01). However, we found no significant difference between either left or right roll inclinations (p=0.51) or between forward and backward pitch inclinations (p=0.59).  On the other hand, we created a predictive model of mental workload (MWL) under ecological conditions. The proposed model is a data-driven machine learning approach that determines the most accurate predictors of mental workload from a set of measured signals. custom paradigm, our algorithm shows good performance (AUC = 0.804 ± 0.086) and produces an interpretable list of the most relevant features for MWL estimation. These results are a first step in determining the perceptual-motor profile of pilots, which could be of great use for their training and longitudinal follow-up.

Rémi Barrois Muller

Title: Spontaneous movement assessment using inertial sensors in infants with severe spinal muscular atrophy after gene therapy.

Abstract: Objective: Spinal muscular atrophy (SMA) is a severe early onset genetic disease with degeneration of motorneurones. Recently, gene therapy (GT) has changed survival and motor prognosis of SMA. Multiple-items motor scores (such as CHOPINTEND) are the gold standard to evaluate motor development in SMA after GT. Inertial measurement units (IMU) are a validated tool to measure movement in toddlers and are a promising tool to complete the traditional motor score assessment. The goal of this pilot study was to provide original IMU data in SMA patients after GT. We propose an original measurement protocol, we show that the 95th percentile of the norm of acceleration (acc_n_95) is a valid parameter to measure spontaneous motor activity and we show its clinical pertinence by testing its association with 30 seconds unaided sitting acquisition.

Methods: All consecutive patients diagnosed with SMA and treated by GT at Necker Hospital between June 2019 and June 2022 were included. Neurological examination, CHOPINTEND and IMU measurements were performed at M0 (before GT), M1, M3, M6, M12, M18 and M24 (24 months after GT). Inertial data was recorded from SMA patients during a 25 minutes spontaneous movement task lying on the back wearing one IMU on each ankle, on each wrist and on each distal part of the arm (just proximal to the elbow). Ten minutes of the test were performed without a toy arch (“toy-“ condition) and 15 minutes with (“toy+”).

Results:  23 patients with severe SMA were included. Acc_n_95 of the upper and the lower limbs showed a good robustness to test conditions (toy- versus toy+) with interclass correlation coefficient of 0.88 and 0.93 respectively. Acc_n_95 increased in all patients with time after GT. Acc_n_95 for upper and lower limbs were well correlated with CHOPINTEND (p-values 4.0e-16 and 6.2e-10 respectively). Acc_n_95 for lower limbs was more strongly associated with the acquisition of 30 seconds unaided sitting than acc_n_95 for upper limbs (area under the curve 0.92 versus 0.90 respectively).

Conclusion: the protocol proposed here is adapted to assess spontaneous movement in infants with severe SMA after GT and was well accepted by patients and families. Spontaneous movements increased with time. Acc_n_95 is a robust parameter. Sensors located on body areas with stronger strength deficit showed higher clinical pertinence.