Biomechanical monitoring and machine learning for the detection of lying postures.

BACKGROUND: Pressure mapping technology has been adapted to monitor over prolonged periods to evaluate pressure ulcer risk in individuals during extended lying postures. However, temporal pressure distribution signals are not currently used to identify posture or mobility. The present study was designed to examine the potential of an automated approach for the detection of a range of static lying postures and corresponding transitions between postures. METHODS: Healthy subjects (n = 19) adopted a range of sagittal and lateral lying postures. Parameters reflecting both the interactions at the support surface and body movements were continuously monitored. Subsequently, the derivative of each signal was examined to identify transitions between postures. Three machine learning algorithms, namely Naïve-Bayes, k-Nearest Neighbors and Support Vector Machine classifiers, were assessed to predict a range of static postures, established with a training model (n = 9) and validated with new input from test data (n = 10). FINDINGS: Results showed that the derivative signals provided a means to detect transitions between postures, with actimetry providing the most distinct signal perturbations. The accuracy in predicting the range of postures from new test data ranged between 82%-100%, 70%-98% and 69%-100% for Naïve-Bayes, k-Nearest Neighbors and Support Vector Machine classifiers, respectively. INTERPRETATION: The present study demonstrated that detection of both static postures and their corresponding transitions was achieved by combining machine learning algorithms with robust parameters from two monitoring systems. This approach has the potential to provide reliable indicators of posture and mobility, to support personalised pressure ulcer prevention strategies.

Personalized modeling for real-time pressure ulcer prevention in sitting posture.

Ischial pressure ulcer is an important risk for every paraplegic person and a major public health issue. Pressure ulcers appear following excessive compression of buttock's soft tissues by bony structures, and particularly in ischial and sacral bones. Current prevention techniques are mainly based on daily skin inspection to spot red patches or injuries. Nevertheless, most pressure ulcers occur internally and are difficult to detect early. Estimating internal strains within soft tissues could help to evaluate the risk of pressure ulcer. A subject-specific biomechanical model could be used to assess internal strains from measured skin surface pressures. However, a realistic 3D non-linear Finite Element buttock model, with different layers of tissue materials for skin, fat and muscles, requires somewhere between minutes and hours to compute, therefore forbidding its use in a real-time daily prevention context. In this article, we propose to optimize these computations by using a reduced order modeling technique (ROM) based on proper orthogonal decompositions of the pressure and strain fields coupled with a machine learning method. ROM allows strains to be evaluated inside the model interactively (i.e. in less than a second) for any pressure field measured below the buttocks. In our case, with only 19 modes of variation of pressure patterns, an error divergence of one percent is observed compared to the full scale simulation for evaluating the strain field. This reduced model could therefore be the first step towards interactive pressure ulcer prevention in a daily set-up.

Biomechanical modeling to prevent ischial pressure ulcers.

With 300,000 paraplegic persons only in France, ischial pressure ulcers represent a major public health issue. They result from the buttocks׳ soft tissues compression by the bony prominences. Unfortunately, the current clinical techniques, with - in the best case - embedded pressure sensor mats, are insufficient to prevent them because most are due to high internal strains which can occur even with low pressures at the skin surface. Therefore, improving prevention requires using a biomechanical model to estimate internal strains from skin surface pressures. However, the buttocks׳ soft tissues׳ stiffness is still unknown. This paper provides a stiffness sensitivity analysis using a finite element model. Different layers with distinct Neo Hookean materials simulate the skin, fat and muscles. With Young moduli in the range [100-500 kPa], [25-35 kPa], and [80-140 kPa] for the skin, fat, and muscles, respectively, maximum internal strains reach realistic 50 to 60% values. The fat and muscle stiffnesses have an important influence on the strain variations, while skin stiffness is less influent. Simulating different sitting postures and changing the muscle thickness also result in a variation in the internal strains.

TexiCare: an innovative embedded device for pressure ulcer prevention. Preliminary results with a paraplegic volunteer.

This paper introduces the recently developed TexiCare device that aims at preventing pressure ulcers for people with spinal cord injury. This embedded device is aimed to be mounted on the user wheelchair. Its sensor is 100% textile and allows the measurement of pressures at the interface between the cushion and the buttocks. It is comfortable, washable and low cost. It is connected to a cigarette-box sized unit that (i) measures the pressures in real time, (ii) estimates the risk for internal over-strains, and (iii) alerts the wheelchair user whenever necessary. The alert method has been defined as a result of a utility/usability/acceptability study conducted with representative end users. It is based on a tactile-visual feedback (via a watch or a smartphone for example): the tactile modality is used to discreetly alarm the person while the visual modality conveys an informative message. In order to evaluate the usability of the TexiCare device, a paraplegic volunteer equipped his wheelchair at home during a six months period. Interestingly, the first results revealed bad habits such as an inadequate posture when watching TV, rare relief maneuvers, and the occurrence of abnormal high pressures.

Finite element speaker-specific face model generation for the study of speech production.

In situations where automatic mesh generation is unsuitable, the finite element (FE) mesh registration technique known as mesh-match-and-repair (MMRep) is an interesting option for quickly creating a subject-specific FE model by fitting a predefined template mesh onto the target organ. The irregular or poor quality elements produced by the elastic deformation are corrected by a 'mesh reparation' procedure ensuring that the desired regularity and quality standards are met. Here, we further extend the MMRep capabilities and demonstrate the possibility of taking into account additional relevant anatomical features. We illustrate this approach with an example of biomechanical model generation of a speaker's face comprising face muscle insertions. While taking advantage of the a priori knowledge about tissues conveyed by the template model, this novel, fast and automatic mesh registration technique makes it possible to achieve greater modelling realism by accurately representing the organ surface as well as inner anatomical or functional structures of interest.

A fast and robust patient specific Finite Element mesh registration technique: application to 60 clinical cases.

Finite Element mesh generation remains an important issue for patient specific biomechanical modeling. While some techniques make automatic mesh generation possible, in most cases, manual mesh generation is preferred for better control over the sub-domain representation, element type, layout and refinement that it provides. Yet, this option is time consuming and not suited for intraoperative situations where model generation and computation time is critical. To overcome this problem we propose a fast and automatic mesh generation technique based on the elastic registration of a generic mesh to the specific target organ in conjunction with element regularity and quality correction. This Mesh-Match-and-Repair (MMRep) approach combines control over the mesh structure along with fast and robust meshing capabilities, even in situations where only partial organ geometry is available. The technique was successfully tested on a database of 5 pre-operatively acquired complete femora CT scans, 5 femoral heads partially digitized at intraoperative stage, and 50 CT volumes of patients' heads. In the latter case, both skin and bone surfaces were taken into account by the mesh registration process in order to model the face muscles and fat layers. The MMRep algorithm succeeded in all 60 cases, yielding for each patient a hex-dominant, Atlas based, Finite Element mesh with submillimetric surface representation accuracy, directly exploitable within a commercial FE software.

Real-time SPECT and 2D ultrasound image registration.

In this paper we present a technique for fully automatic, real-time 3D SPECT (Single Photon Emitting Computed Tomography) and 2D ultrasound image registration. We use this technique in the context of kidney lesion diagnosis. Our registration algorithm allows a physician to perform an ultrasound exam after a SPECT image has been acquired and see in real time the registration of both modalities. An automatic segmentation algorithm has been implemented in order to display in 3D the positions of the acquired US images with respect to the organs.