The rapidly growing volume of electrophysiological signals has been generated for clinical research in neurological disorders. European Data Format (EDF) is a standard format for storing electrophysiological signals. However, the bottleneck of existing signal analysis tools for handling large-scale datasets is the sequential way of loading large EDF files before performing an analysis. To overcome this, we develop Hadoop-EDF, a distributed signal processing tool to load EDF data in a parallel manner using Hadoop MapReduce. Hadoop-EDF uses a robust data partition algorithm making EDF data parallel processable. We evaluate Hadoop-EDF’s scalability and performance by leveraging two datasets from the National Sleep Research Resource and running experiments on Amazon Web Service clusters. The performance of Hadoop-EDF on a 20-node cluster improves 27 times and 47 times than sequential processing of 200 small-size files and 200 large-size files, respectively. The results demonstrate that Hadoop-EDF is more suitable and effective in processing large EDF files.

De nos jours, le domaine biomédical est en pleine croissance avec le développement de dispositifs thérapeutiques innovants, bas coût, pour le diagnostic, le traitement ou la prévention de maladies chroniques ou cardiovasculaires. Ces dernières années ont connu l’émergence des polymères semi-conducteurs, alternative intéressante aux matériaux inorganiques, présentant des propriétés uniques de conduction ionique et électronique. Tout d’abord, j’ai axé mes travaux de recherche sur le développement et l’optimisation d’une encre conductrice à base de PEDOT:PSS, parfait candidat comme matériau, pour la transduction des signaux biologiques en signaux électriques, compatible avec le process jet d’encre, pour la réalisation de dispositifs imprimés. Puis mes travaux se sont orientés vers la conception et l’étude d’électrodes imprimées sur supports papiers, tatous et textiles permettant des enregistrements long termes d’électrocardiogrammes (ECG) ou électromyogrammes (EMG), présentant des performances similaires aux électrodes commerciales, utilisant un système d’acquisition spécifique pour la mesure d’activités électriques de tissus musculaires. Puis dans un second temps, je me suis penchée sur l’impression sur support papier, de transistors organiques électrochimiques (OECTs) fonctionnalisés, afin de permettre la détection d’éléments biologiques ou chimiques comme l’alcool. Ces travaux proposent une nouvelle voie pour la conception de dispositifs innovants biomédicaux à bas couts, imprimés, permettant la personnalisation des produits pouvant être intégrés dans des dispositifs biomédicaux portables ou dans des vêtements « intelligents ». / With the evolution of microelectronics industry and their direct implementation in the biomedical arena, innovative tools and technologies have come to the fore enabling more reliable and cost-effective treatment. In this thesis I focus on the integration of the conducting polymer PEDOT:PSS with printing technologies toward the realization of performant biomedical devices. In the first part, I focus on the optimization of the conducting ink formulation. Following, I emphasize on the fabrication of inkjet printed PEDOT:PSS based biopotential electrodes on a wide variety of substrates (i.e., paper, textiles, tattoo paper) for use in electrophysiological applications such as electrocardiography (ECG) and electromyography (EMG). Printed electrodes on paper and printed wearable electrodes were fabricated and investigated for long-term ECG recordings. Then, conformable printed tattoo electrodes were fabricated to detect the biceps activity during muscle contraction and the conventional wiring was replaced by a simple contact between the tattoo and a similarly ink-jet printed textile electrode.In the last part, I present the potentiality of inkjet printing method for the realization of organic electrochemical transistor (OECTs) as high performing biomedical devices. A disposable breathalyzer comprised of a printed OECT and modified with alcohol dehydrogenase was used for the direct alcohol detection in breath, enabling future integration with wearable devices for real-time health monitoring. Their compatibility with printing technologies allows the realization of low-cost and large area electronic devices, toward next-generation fully integrated smart biomedical devices.

Driver sleepiness is a cause for crashes and it is estimated that 3.9 to 33 % of all crashes might be related to sleepiness at the wheel. It is desirable to get an objective measurement of driver sleepiness for reduced sensitivity to subjective variations. Using deep learning for classification of driver sleepiness could be a step toward this objective. In this master thesis, deep learning was used for investigating classification of electrophysiological data, electroencephalogram (EEG) and electrooculogram (EOG), from drivers into levels of sleepiness. The EOG reflects eye position and EEG reflects brain activity. Initially, the intention was to include electrocardiogram (ECG), which reflects heart activity, in the research but this data were later excluded. Both raw time series data and data transformed into time-frequency domain representations were fed into the developed neural networks and for comparison manually extracted features were used in a shallow neural network architecture. Investigation of using EOG and EEG data separately as input was performed as well as a combination as input. The data were labeled using the Karolinska Sleepiness Scale, and the scale was divided into two labels "fatigue" and "alert" for binary classification or in five labels for comparison of classification and regression. The effect of example length was investigated using 150 seconds, 60 seconds and 30 seconds data. Different variations of the main network architecture were used depending on the data representation and the best result was given when using a combination of a convolutional neural network (CNN) and a long short-term memory (LSTM) network with time distributed 150 seconds EOG data as input. The accuracy was in this case 80.4 % and the majority of both alert and fatigue epochs were classified correctly with 85.7 % and 66.7 % respectively. Using the optimal threshold from the created receiver operating characteristics (ROC) curve resulted in a more balanced classifier with 76.3 % correctly classified alert examples and 79.2 % correctly classified fatigue examples. The results from the EEG data, both in terms of accuracy and distribution of correctly classified examples, were shown to be less promising compared to EOG data. Combining EOG and EEG signals was shown to slightly increase the proportion of correctly classified fatigue examples. However, more promising results were obtained when balancing the classifier for solely EOG signals. The overall result from this project shows that there are patterns in the data connected to sleepiness that the neural network can find which makes further work on applying deep learning to the area of driver sleepiness interesting.

<p>In modern medicine, the measurement of electrophysiological signals play a key role in health monitoring and diagnostics. Electrical activity originating from our nerve and muscle cells conveys real-time information about our current health state. The two most common and actively used techniques for measuring such signals are electrocardiography (ECG) and electroencephalography (EEG).</p><p>These signals are very weak, reaching from a few millivolts down to tens of microvolts in amplitude, and have the majority of the power located at very low frequencies, from below 1 Hz up to 40 Hz. These characteristics sets very tough requirements on the electrical circuit designs used to measure them. Usually, measurement is performed by attaching electrodes with direct contact to the skin using an adhesive, conductive gel to fixate them. This method requires a clinical environment and is time consuming, tedious and may cause the patient discomfort.</p><p>This thesis investigates another method for such measurements; by using a non-contact, capacitively coupled sensor, many of these shortcomings can be overcome. While this method relieves some problems, it also introduces several design difficulties such as: circuit noise, extremely high input impedance and interference. A capacitively coupled sensor was created using the bottom layer of a printed circuit board (PCB) as a capacitor plate and placing it against the signal source, that acts as the opposite capacitor plate. The PCB solder mask layer and any air in between the two acts as the insulator to create a full capacitor. The signal picked up by this sensor was then amplified by 60 dB with a high input impedance amplifier circuit and further conditioned through filtering.</p><p>Two measurements were made of the same circuit, but with different input impedances; one with 10 MΩ and one with 10 GΩ input impedance. Additional filtering was designed to combat interference from the main power lines at 50 Hz and 150 Hz that was discovered during initial measurements. The circuits were characterized with their transfer functions, and the ability to amplify a very low-level, low frequency input signal. The results of these measurements show that high input impedance is of critical importance for the functionality of the sensor and that an input impedance of 10 GΩ is sufficient to produce a signal-to-noise ratio (SNR) of 9.7 dB after digital filtering with an input signal of 25 μV at 10 Hz.</p>

In modern medicine, the measurement of electrophysiological signals play a key role in health monitoring and diagnostics. Electrical activity originating from our nerve and muscle cells conveys real-time information about our current health state. The two most common and actively used techniques for measuring such signals are electrocardiography (ECG) and electroencephalography (EEG). These signals are very weak, reaching from a few millivolts down to tens of microvolts in amplitude, and have the majority of the power located at very low frequencies, from below 1 Hz up to 40 Hz. These characteristics sets very tough requirements on the electrical circuit designs used to measure them. Usually, measurement is performed by attaching electrodes with direct contact to the skin using an adhesive, conductive gel to fixate them. This method requires a clinical environment and is time consuming, tedious and may cause the patient discomfort. This thesis investigates another method for such measurements; by using a non-contact, capacitively coupled sensor, many of these shortcomings can be overcome. While this method relieves some problems, it also introduces several design difficulties such as: circuit noise, extremely high input impedance and interference. A capacitively coupled sensor was created using the bottom layer of a printed circuit board (PCB) as a capacitor plate and placing it against the signal source, that acts as the opposite capacitor plate. The PCB solder mask layer and any air in between the two acts as the insulator to create a full capacitor. The signal picked up by this sensor was then amplified by 60 dB with a high input impedance amplifier circuit and further conditioned through filtering. Two measurements were made of the same circuit, but with different input impedances; one with 10 MΩ and one with 10 GΩ input impedance. Additional filtering was designed to combat interference from the main power lines at 50 Hz and 150 Hz that was discovered during initial measurements. The circuits were characterized with their transfer functions, and the ability to amplify a very low-level, low frequency input signal. The results of these measurements show that high input impedance is of critical importance for the functionality of the sensor and that an input impedance of 10 GΩ is sufficient to produce a signal-to-noise ratio (SNR) of 9.7 dB after digital filtering with an input signal of 25 μV at 10 Hz.