13:00
Motor Control, Neuro Control & Patient Models IV
13:00
15 mins
|
MEASUREMENT OF NEUROMECHANICAL ANKLE PARAMETERS IN CHILDREN WITH CEREBRAL PALSY
Lizeth Sloot, Marjolein van der Krogt, Karin de Gooijer, Carel Meskers, Jurriaan de Groot, Annemieke Buizer, Jules Becher, Erwin de Vlugt, Jaap Harlaar
Abstract: Spastic Cerebral Palsy (CP) is characterized by increased joint stiffness, caused by a mix of increased stretch reflex activity and muscle tone, as well as tissue stiffness. Since treatment depends on the specific cause, objective quantification of the ankle neuromechanical parameters would contribute to patient specific treatment. Previously developed instrumented assessment (IA) [1,2] was extended to include background muscle tone and to differentiate between the three major lower leg muscles. We evaluated the ability of the extended IA to discriminate between children with spastic CP and controls.
The most affected leg of 35 CP children (11.1±3.3 yr, GMFCS 1-3) and 35 controls (10.2±2.7 yr) was measured. A motor driven footplate equipped with a force transducer and potentiometer applied 2 passive slow (5 º/s) and fast (100 º/s) ramp-and-hold rotations around the ankle joint, at two different knee angles (20º and 70º) to discriminate between the triceps muscles. Ankle angle and EMG of the gastrocnemius (GAS), soleus (SOL) and tibialis anterior (TIB) muscles were used to optimize a nonlinear neuromuscular model to match the measured ankle torque. Tissue stiffness and viscosity were based on the slow trials and taken at the highest plantar (40º) and dorsiflexion (5º) angle reached by all subjects. Root-mean-squares of baseline muscle tone and reflex torque were taken from the fast trials. Non-parametric tests with Bonferroni correction were performed to assess the difference between CP and controls.
In CP, stiffness was 1.7 times increased in SOL (p=0.04), showed a trend of a 1.8 increase in GAS (p=0.13) and was 0.63 times decreased in TA (p=0.03). Reflex activity was 3.5 and 3.1 times increased in CP for SOL (p<0.01) and GAS (p<0.01) respectively. Background muscle activity was 1.7 times increased in SOL (p<0.01). Variances were generally larger for CP and ratios between stiffness and reflex torque differed considerably between patients.
The IA was able to discriminate between CP patients and controls. In CP, reflex torque and muscle tone were found to be increased in the triceps muscles, which are often treated for spasticity. Stiffness only showed an increase for SOL, although a difference also emerged for GAS when choosing a higher dorsiflexion angle, at the expense of decreasing the group of subject. The large variances and ratio differences in CP indicates that instrumented assessments could allow for subject specific therapy selection.
|
13:15
15 mins
|
BIOMECHANICS BASED ANALYSIS OF SLEEP
Tim Willemen, Bart Haex, Jos Van der Sloten, Sabine Van Huffel
Abstract: Current clinical standards to assess sleep and its disorders lack either accuracy or user-friendliness. They are therefore difficult to use in cost-effective population-wide screening or long-term objective follow-up after diagnosis. In order to fill this gap, the use of off-body detected cardiorespiratory information was evaluated for sleep stage discrimination and detection of apneic breathing.
Sleep stage classification was performed based on cardio-respiratory measures extracted from electrocardiography (ECG) and breathing belts. Movement measures were extracted from a continuously monitored mattress indentation pattern. Feature selection was performed using Mutual Information filter methods, and Cohen’s kappa wrapper methods. Both Support Vector Machines as well as Linear Discriminant Analysis were used as classification methods.
A healthy population dataset (part of a private database) contained a total of 85 nights, with annotated sleep stages. A first apneic population dataset (the Dublin Sleep Apnea database, publicly available at Physionet) contained a total of 25 nights, with both annotated sleep stages as well as apneic events. No movement signal was included in this database. A second apneic population dataset (the Apnea-ECG database, also publicly available at Physionet) contained a total of 70 nights, with annotated apneic events. Only an ECG signal was included in this database.
In distinguishing between Wake, REM and NREM, an overall agreement of 81% was achieved on the healthy dataset, and 70% on the first apneic dataset. The difference is mainly explained by the significant impact of apneic events on the cardiorespiratory system, independent of the sleep stage. Apneic breathing could be distinguished from healthy breathing with an agreement of 76% on the first apneic dataset and 90% on the second apneic dataset. The difference is mainly explained by the larger apnea-hypopnea ratio in the second database (almost 40 times higher!), since apneic events are much easier to detect compared to their less severe hypopneic variants.
In order to assess the reliability of off-body detected cardiorespiratory information, fifteen subjects lied down for 10 minutes on the pressure-based ballistocardiographic (BCG) setup, while changing posture every 2.5 minutes (supine, left lateral, prone, right lateral). Heart beats were detected using the Pan-Tompkins algorithm, adapted for use on the BCG, including a template-based realignment of heart beat location estimates. An interbeat-interval agreement of 97% was achieved between heart beats detected in the BCG and those detected in the reference ECG. Signal quality was the highest in supine, and the lowest in right lateral.
|
13:30
15 mins
|
A PROSPECTIVE FMRI-BASED TECHNIQUE FOR LOCALIZING THE EPILEPTOGENIC ZONE IN PRESURGICAL EVALUATION OF EPILEPSY
Borbála Hunyadi, Simon Tousseyn, Patrick Dupont, Sabine Van Huffel, Maarten De Vos, Wim Van Paesschen
Abstract: EEG-correlated fMRI analysis has proven to be useful in localizing regions of BOLD activation related to epileptic activity. However, as EEG does not always provide reliable information, purely fMRI-based data-driven techniques are invaluable. Recently, we have shown that independent component analysis (ICA) can extract sources related to the epileptic network even in such EEG-negative cases [1]. In order to utilize this concept in clinical practice in a prospective manner, we developed an automatic technique for selecting epileptic sources [2]. The proposed approach applies a cascade of two classifiers. In the first step artifact related sources are discarded. In the second step the sources are characterized by four discriminative features and epileptic sources are selected from among other BOLD-related components. Tested on a patient population of 10 EEG-positive cases, our technique provided concordant sources with the EEG-correlated fMRI activation maps or with the seizure onset zone in 71% of the cases. Nevertheless, the proposed method is especially beneficial in cases where no interictal discharges are visible in the EEG. Therefore, in [3] we validated the proposed method in a group of 18 EEG-negative patients and 13 healthy controls, in order to assess the behavior of the method in the absence of epileptic activity. Our technique selected an epileptic source in 4 cases. All 4 maps were correct and unequivocally indicated a region in the brain corresponding to the epileptogenic zone. Moreover, the method made a selection in only 1 out of 13 controls. The significance of our technique lies within the fact that it is very specific and fully automated. Therefore, we conclude that it can be used prospectively in the presurgical evaluation of epilepsy.
|
13:45
15 mins
|
DERIVATION OF EPIDERMAL NERVE FIBER DENSITY FROM THE RESPONSE TO LASER STIMULI APPLIED WITH A SPATIAL FILTER: PILOT STUDY IN VOLUNTEERS
Renske Hoeben, Imre Krabbenbos, Eduard Boezeman, Christiaan van Swol
Abstract: One of the most commonly reported symptoms in sarcoidosis patients is chronic pain. Nerve conduction studies and other standard diagnostic methods often fail to detect objective evidence of neurological disease in these patients. In this study we would like to explore a different method to detect changes in the nervous system in sarcoidosis patients.
Pain is perceived through two different pathways involving either Aδ- fibers or C-fibers. Aδ-fibers are thinly myelinated fibers responsible for a quick, sharp pain (e.g., when you burn your fingers) and C-fibers are unmyelinated fibers that relate to a slower, but lingering response (e.g., a tingling sensation after someone slapped you). Several studies suggest that Aδ- and C-fiber responses can be used to objectively determine nerve disease involving the Aδ- and/or C-fiber systems [1].
Previous studies primarily focused on response times between stimulus and perception, but epidermal nerve fiber density of both Aδ- and C-fibers might be another predictor of a neurological disease. Epidermal nerve fiber density is normally determined by a skin biopsy, which is a painful procedure and cannot discriminate between Aδ- and C-fibers. The goal of this study is to investigate whether it is feasible to determine the epidermal nerve fiber density (Aδ and C) using a non invasive technique.
We will use differences in heat activation threshold and epidermal distribution density to selectively activate Aδ- and C-fibers. Mouraux et al. showed that a constant increase in stimulus surface area leads to higher laser detection rates [2]. We will use this theory to relate the detection of stimulation to the individual’s nerve fiber density.
The test-subjects will be subjected to stimuli from a 980nm diode laser (Ceralas D15, Biolitec AG). To selectively stimulate Aδ- and C-fibers two different laser settings are used: for Aδ- 20ms at 2.5W and for C-fibers 50 ms at 0.5W. A spatial filter is used to apply the stimuli with different surface areas. The spatial filter contains 50 holes with varying diameters (0.4, 0.6, 1, 2, and 2.5 mm). The test subject receives 100 stimuli in total, and the detection rate of these stimuli will be used to derive the test subject’s nerve fiber density.
The results of this study are expected at the end of 2014 and will be discussed at the conference. This pilot study will serve as a basis for an upcoming clinical investigation in the differential involvement of Aδ- and C-fibers in small fiber neuropathies.
|
14:00
15 mins
|
MEASURING PATIENT AND INFORMATION FLOW IN SURGICAL DAY CARE USING RFID TECHNOLOGY
Linda Wauben, Annetje Guédon, John van den Dobbelsteen
Abstract: Due to the rising number of patients, hospitals are urged to improve efficiency and focus on patient centred care. Radio Frequency IDentification (RFID) technology can be used to reach these goals. The aim of this study was to measure wait times for patients undergoing eye surgery during surgical day care, and characterize the current information flow between staff and patients and between staff from different departments and their future wishes.
This study was conducted at the Rotterdam Eye Hospital. Active RFID technology was used to automatically track adult patients’ location during their entire hospital stay. Active RFID tags (pulse rate 0.8, frequency 433.92 MHz, power 1mW, weight 24 gram) were attached to the patients’ identification wristbands and were tracked by readers (GW3D, RePoint, the Netherlands) that were placed at eight locations along the surgical trajectory. Length of hospital stay and wait times per phase were generated. Patients’ accompanying persons and ward nurses were also interviewed concerning the current information flow and their wishes for the future.
In total, 622 patients were included (405 GA general anaesthesia patients, n=217 LTA local or topical anaesthesia patients). GA patients spent on average 7h01 in hospital and LTA patients on average 4h17. Patients had to wait in each phase (see table). The total percentage of wait and wait-recovery time during the entire hospital stay ranged for GA patients from 0-87.0% with an average of 68.2% and for LTA patients between 20.8-85.7% with an average of 64%.
Thirty accompanying persons were interviewed and although most accompanying persons received information on the duration of the surgery and the arrival time at the postoperative ward (n=18), most accompanying persons would also like to be informed about the progress in the surgical centre (n=19). The nine ward nurses indicated that most problems concerned late phone calls concerning bringing patients to the holding and picking up patients from the recovery. In the future, almost all ward nurses would like to be informed (via an information system) when the patient is ready for the next phase.
RFID technology is a practical tool to measure patient flow automatically and real-time for patients undergoing eye surgery during surgical day care. Based on this study a “Patient Tracking System” was designed and implemented (nine months ago) together with the ward nurses, patients and accompanying persons to communicate the patient flow real-time.
|
14:15
15 mins
|
INSTRUMENTED SPASTICITY ASSESSMENT IN CHILDREN WITH CP: MOTOR-DRIVEN VERSUS MANUAL TESTING
Lizeth Sloot, Lynn Bar-On, Marjolein van der Krogt, Annemieke Buizer, Guy Molenaers, Erwin Aertbelien, Jules Becher, Erwin de Vlugt, Kaat Desloovere, Jaap Harlaar
Abstract: Current clinical assessment of increased joint resistance (stiffness and spasticity), is based on subjective manual evaluation of the resistance to slow and fast movements. Recently, it has been shown that instrumented manual testing is superior in terms of objectivity, precision and resolution [1]. However, since spasticity is considered to be both force- and velocity- dependent, robotic manipulations have also been developed [2]. This study compared the outcomes of motor-driven (MO) to manual (MA) instrumented spasticity assessment in children with spastic Cerebral Palsy (CP).
The most affected leg of nine children with CP (10.5±2.9 yr, GMFCS I-III) was measured. In MO, a motor driven footplate equipped with a force transducer and potentiometer applied 2 passive slow (5 º/s) and fast (100 º/s) ramp-and-hold rotations around the ankle joint. In MA, rotations of comparable speed were applied using a foot orthotic attached to a hand-held force transducer and equipped with 2 inertial sensors [1]. EMG from the medial and lateral gastrocnemius, soleus and tibialis anterior was measured. Presence of spasticity (yes or no) was subjectively scored by one rater for each fast trial for both MO and MA. Parameters related to range of motion (ROM), maximum acceleration (AccMax), average RMS-EMG and work were compared between MO and MA using Spearman’s correlation coefficients.
There was a 71% agreement between MO and MA for the presence of spasticity. On average, triceps surae activation was higher in MO than MA during fast, but not during slow stretches. Significant, high correlations were found for ROM (r=0.91) and work (r=0.80) in the slow movement, but not for the other parameters in slow or fast movement.
Outcome spasticity parameters differed between the two conditions. We hypothesize that the different movement profiles (high initial acceleration in MO versus intermittent acceleration in MA) accounted for the disparate central nervous system response to passive stretch. Further research should establish which type of imposed muscle stretch best represents active tasks, such as walking. This research contributes to the search for an instrumented alternative that will facilitates improved treatment planning and outcome evaluation in spastic patients.
|
|
