en

User-friendly assessment tool for motor function (or dysfunction)

Interventions & Care

Neurocontrol-on-the-move

Achieving better diagnosis, higher productivity and lower cost – those are the aims of a group developing supporting technology for disorders affecting motor function. LUMC’s Technology In Motion lab shows that these aims can be attained, in part by using new measurement systems using Microsoft KinectTM.

One example of such a system is the Interactive Walkway for assessing gait. Another (developed by TU Delft Vision Lab) is custom software for fine-grained analysis of movement disorders.

The Interactive Walkway is a user-friendly and broadly applicable assessment tool for motor function (or dysfunction), which allows for unobtrusive automated quantification and classification of gait and gait adaptability, which can then be tailored to the individual patient.Dr. Linda van Schaik-Bank, Human movement scientist

At the moment, expert opinion and wired marker-based measurements are in many cases the basis for evaluation of motor function (or dysfunction) in patients. But expert opinion, however valuable it may be, lacks uniformity and objectivity. In order to assess characteristics of gait, patients are traditionally covered with a number of adhesive markers to record movement, but this is as inconvenient for the patient as its preparation is time-consuming for professionals. Moreover, the clinical lab environment does not resemble patient’s natural, daily life experience. Clinical tests are often conducted in a deprived setting, disregarding possible deficiencies in the ability to adjust movements in response to challenges in the environment, which forms an essential part of functioning in daily life. Isn’t there a better way to establish how the patient moves, what movements deviate and what the origin of this deviation might be?”

Markerless

“Innovative technologies can address both aspects,” says Dr. Van Schaik-Bank. “They allow doctors to assess the severity and stage of the patient’s condition uniformly and objectively.” This is the aim of Technology In Motion (TIM). It forms part of the Neurocontrol Programme within the Innovative Medical Device Initiative (IMDI). The purpose of this initiative is to develop tools and technologies for assessing neuroplasticity in diseases such as stroke, spinal cord injury and Parkinson’s Disease. The latter is the focus of LUMC. An example of such an assessment tool is the haptic robot in the TIM laboratory. This robot was developed in an earlier project (called TREND) at TU Delft. Patients hold a handle, to which pulses of distorting vibrations on specific frequencies are introduced. Measurement of reflex reactions to the handle and analysing the electromyography (EMG) that scans muscle activity, provides information on dystonia, the location and magnitude of the problem, reaction speed and power feedback. “An important feature of this haptic robot,” says Dr Van Schaik-Bank, “is that it enables markerless measurements of arm and hand movement. These measurements are the basis that we use for translation to underlying mechanisms.”

Interactive Walkway

Flagship of the TIM laboratory at LUMC is a markerless solution for evaluating the adaptability of gait: the Interactive Walkway (IWW), It is based on Microsoft Kinect.TM With its combination of RGB colour-imaging and infrared depth projection, Kinect has inspired a lot of professional applications. It was designed by human movement scientist Melvyn Roerdink and developer Bert Coolen, of VU University’s Department of Human Movement Sciences. It combines four Kinect sensors and covers an eight-metre walkway for markerless assessment of gait, with a projector to enrich the patient’s environment with visual content.

The IWW is the basis of PhD research being conducted by Daphne Geerse. Among other things, her task is to validate the new method in healthy individuals and in various patient populations. First, the patient´s gait parameters will be measured during normal unperturbed walking (e.g., speed, cadence, step length and step width). Then, the information obtained is used for assessing gait adaptability. Targets and obstacles are projected on the floor in front of the patient. How does the patient cope with this situation? During this stage, the patient may show (remaining) adaptive capabilities, which may reveal whether there is an increased risk of fall incidents.

“As the examples of the haptic robot and the IWW make clear,” says Van Schaik-Bank, “you can learn most about a system when you apply perturbations to it. This not only provides information on the functional integrity of the systems involved in motor control, but also enables us to better tune screening for fall-risk to the most important causes and circumstances of falls in daily life. A big plus is the generic nature of the IWW and the movement analysis behind it. This can be deployed for motor dysfunction in general: it can prove its value in various disorders that affect motor function and always enables fine-tuning to the individual level.”

Artificial Intelligence

As the standard KinectTM software is not designed for medical applications, specific software is under development. This is where Jan van Gemert from the TU Delft Vision Lab comes in: “The standard software only allows for representation of a person’s skeleton in rough outline. Although these estimated body points proved very useful for the quantification of gross arm movements and global gait parameters, identification of subtle deviances in movement patterns or accurate quantification of hand and finger movements requires high resolution, accurate depth measurement and 3D representation of the individual. For instance, we focus on zooming in on the minuscule movements, which can be most revealing about the patient’s condition.” A next step, says Van Gemert, would be to introduce pattern recognition and artificial intelligence into the process, and to use the computer for supporting the diagnosis on the basis of objective measurement of frequency, amplitude and phase. “The latest deep-learning techniques in artificial intelligence are rivaling human performance in image analysis,” he says.

Good collaboration

Bank speaks of very good collaboration within the consortium. “The key is to understand each other’s professional language as well as possible.” Van Gemert, just like Van Schaik-Bank, is firmly convinced of the added value of the multidisciplinary approach. “A case like this one enables solutions that could not be imagined beforehand, and allows for new directions that would not emerge while working on your own. It’s very stimulating to work on practical solutions like these.”

See also https://tim.lumc.nl

Interview by: Leendert van der Ent