Comnets.uni-bremen.de

Indicating Motor Symptoms in PD Patients Using Abstract—Here the research focus of a dissertation is pre- No convenient and effective way to measure dopamine sented. It focuses on evaluation and improvement of algorithms levels within the basal ganglia circuit and adjust medication for recognizing Parkinson’s Disease (PD) motor symptoms in intake with accordance to dopamine levels has been found time series data. PD is a disorder of the central nervous system yet. Instead the presence and absence of symptoms is used to resulting in a loss of motor function, increased rigidity and evaluate the effectiveness of a treatment. Generally speaking, slowness. Through application of artificial intelligence (AI)-based the longer the patient doesn’t show symptoms the better. It techniques, the occurrence of symptoms such as tremor or is a common problem for neurologists to find a dose of bradykinesia can be indicated in time series generated by sensorsworn on the patient’s body. Those affected by PD bear a great medication that optimizes the treatment for a particular patient burden and have to cope with a rather reduced quality of life.
in a way that doesn’t provoke any unnecessary side-effects Minimizing false negatives and false positives enables a better but keeps the patient fluid and in control of their movements.
treatment of people with PD as the applied drug dosage can be Projects like HELP [6] and REMPARK [7] try to automate set in accordance with currently apparent symptoms (rather than this process of manual medication adjustment. They rely on a global dosage as current treatments employ). The paper gives the proper recognition of PD (motor) symptoms in order to an overview for a doctoral colloquium to discuss the intended estimate the required drug level. Thus, the research focus is here the comparison and improvement of state of the artsymptom indication algorithms as well as development of new approaches. The research hypothesis is that combining suchalgorithms results in an improved indication of PD symptoms.
PD is a chronic, progressive, neurodegenerative disorder These efforts could provide a foundation for better monitoring [1], [2] which is generally characterized by gradual loss of systems and therefore also for better treatment of PD patients function (e.g. slowness and rigidity). Even though it can and consequently reducing the burden for everyone involved manifest itself at any age, PD is usually attributed to elderly (i.e. patients, caretakers, relatives, friends, etc.).
subgroups of the population. The cardinal symptoms arebradykinesia, rigidity, tremor and postural instability [1], [2], [3]. Among many other symptoms, these symptoms result froma dopamine (a neurotransmitter used to control movement) In the early 19th century, James Parkinson first described deficiency in the substantia nigra (a part of the brain related in “An essay on the shaking palsy”[8] six cases showing the basal ganglia circuit). Usually by the time of diagnosis, (motor) symptoms such as a shaking hand or slowness of a great number of dopamine-producing neurons have already movement. Even after 200 years of advancements in medical diminished [1]. The authors of [4] presume a long pre-clinical technologies, the cause of PD remains unknown [9], [1], [4], phase (i.e. 10 - 20 years), in which the symptoms remain [10]. As to what triggers the disease, it can only be speculated.
Researchers are investigating various possibilities includingviruses, environmental factors, aging and genetic causes [11], Degeneration of dopamine-producing neurons continues [12], [1], but no definitive answer can be given at this point in with progression of the disorder and renders those suffering time. Considering the diversity of the disease (e.g. genetic and from PD more and more dependent on assistance by caretakers, non-genetic origin) and it’s large variety of symptoms, there family and friends. Unfortunately, there is no cure but with are educated speculations that PD may in fact not be a single proper treatment much of the lost quality of life can be reclaimed. These treatments aim at slowing the progressionof the disease, focus on symptomatic relief and attempt to Like Alzheimer’s, dementia and chronic bronchitis, PD lift the enormous burden of PD. However, this is not an easy is usually attributed to people in their senior years (60 and task due to uncertainties in diagnostic procedures (a definitive above). As the population is aging, the number of cases and diagnosis requires autopsy [1]), a lack of methods for measur- burden of PD is expected to increase [13, p. 36]. The World ing dopamine levels and unknown origin of the illness. Thus Health Organization (WHO) estimates that around 5.2 million doctors have to revert to clinical procedures and guidelines people were suffering from PD worldwide in 2004 [14].
which leave room for interpretation and subjective judgment.
Depending on the estimating organization, Europe inhabited It is not uncommon to find misdiagnosed PD patients [5].
1.2 [13] - 2.0 [14] million of them in the same year.
As the authors of “Cost of disorders of the brain in Europe state) and in which they show symptoms because the effects 2010” [13] note, a patient bears a yearly cost of about 11.200 of the medication have “worn off” (“OFF” state). Keeping the Euro, which adds up to total cost of almost 14.000 million patient in the ON state is desirable. Depending on the stage of Euro within Europe alone (2010). To set this into perspective, the disease, patients will fluctuate between ON and OFF states Germany inhabited the second largest number of people with several times during the day (e.g. medium-advanced patients Parkinson’s in the same year (highest number of subjects with PD was in Italy with close to 240.000 people, compared to In an early stage of the disease, these fluctuations are pre- about 220.000 people in Germany). They are followed by dictable and can almost entirely be avoided by a proper timing France and United Kingdom with 190.000 and 110.000 people of medication intake and it’s dosage. However, as the disease respectively. In Germany, the total cost of PD was roughly progresses the desired effects of levodopa “wear off”. Thus 2.800 million Euro whereas on average each subject bare a the effectiveness of medication has a tendency to decrease and causes the symptoms to reappear earlier than before [12].
In the Global Burden of Disease (GBD) study, the WHO With continued progression of the disease (degeneration of rated PD to be on the same disability level as: amputated arm, dopamine neurons) and reduced effectiveness of medication, congestive heart failure, deafness, drug dependence and tuber- the only working option is to increase the dose in the hope culosis [14, p. 33]. Considering neuropsychiatric disorders, PD of rendering the patient mobile. Thus the dose needs to be is after Alzheimer and Epilepsy the third most frequent cause increased from time to time. This also means that fluctuations of death in the world [14, p. 56]. But PD is a great burden, not between periods of mobility and immobility happen more just for people suffering from the disease but also for those frequently as the disease progresses. As mentioned above, at being indirectly affected (i.e. relatives and caretakers) due to first these fluctuations are predictable. However, in later stages time consuming assistance. In an advanced stage of the disease of the disease, the fluctuations become less deterministic and and without proper treatment, patients are no longer capable totally unpredictable eventually (ON-OFF phenomenon) [16].
of taking care for themselves. They are highly dependent ontheir caretakers and relatives.
The fluctuations between periods in which individuals show almost no motor symptoms and periods in which motor Various tools and techniques exist to help and assist PD symptoms are present are a major problem for people with patients. Unfortunately, the loss of dopamine cannot directly Parkinson’s. Especially in later stages of the disease where be monitored as it can be done with other diseases such as these fluctuations are less deterministic. It is a considerable diabetes. There, blood sugar levels can be observed and a burden for them to fear a sudden fluctuation while they are system can adjust the amount of required medication based in a public place or somewhere with their friends / family.
on this observation. With PD this is a difficult and invasive Imagine this happening while crossing the street, drawing task because measuring dopamine levels requires direct access closer to the bus / train stop or in a restaurant. Due to social to the brain stem. Thus many of the PD cases can only be complications of their fluctuations, PD patients tend to reduce confirmed post-mortem through an autopsy [15].
their social interaction and prefer to stay in their apartment.
A large number of symptoms have been shown by people Unfortunately, this typically results in less movement and may with Parkinson’s [2], [10]. The most visible and easily no- have further implications on the illness or provoke different ticeable symptoms are related to motor functions. However, the quality of life is also affected by non-motor symptomslike depression, sleep disorder, cognitive / neurobehavioral abnormalities, autonomic and gastrointestinal dysfunction [2], Much research has been done with regard to PD. Innumer- [10], [3]. As the disease progresses, patient’s symptoms change able papers have been published with a focus on biological, and fluctuate (i.e. some symptoms simply disappear, while chemical and genetic aspects of the disorder. However, it others (re-) appear), creating a unique symptomatic history was only until the 1950/60s that PD treatment improved for individual patients. Unfortunately, in an advanced stage significantly through the discovery of dopamine and medica- of the disease further (drug-induced) symptoms may become tions like levodopa. Nonetheless, a number of contributions apparent. Dyskinesia is one of these symptoms and results have been produced that originate from fields like computer from a lengthy pharmacological treatment (i.e. several years).
science and AI. These publications reveal a great number of It manifests itself as an involuntary movement of entire body techniques for automatically indicating the presence of PD parts (e.g. rhythmical moving of upper body).
motor symptoms. Various AI-based methods such as neural In order to reduce the symptoms and compensate for the networks [17], [18], [19], [20], [21], [22], [23], [24], hidden loss of dopamine, patients usually take medication such as markov model (HMM)s [25] and support vector machines [20], levodopa. Most of the prescribed medication is given in pulses [26], [24]. Depending on the symptom and utilized sensors, (e.g. pill taken every few hours) and not continuously (e.g. a various features are calculated (e.g. entropy [20], [26], [24], pump continuously releasing a constant amount of drug). As [27], spectral or fractal features [28], [29], [30], [19], [31], the disease progresses, medication decreases in effectiveness [32], [33], [34], [35], [36]) are known to be used in this and “wears off” which results in the reappearance of symptoms context. Over time, sensor signals are analyzed and compared before the next dose is taken. Furthermore, the so-called (or set in relationship) to known samples of each symptom in therapeutic window, in which the medication produces the order to recognize them. No matter whether these AI methods desired (positive) effects narrows with progression of the dis- are continuous or window-based, they all can be viewed as ease. Consequentially, moderate and advanced patients cycle algorithms that are applicable for time series analysis and data between states in which they can move almost normally (“ON” mining techniques. However, much of the literature presents algorithms for recognizing individual symptoms (e.g. [37], server and web-portal. The server analyzes and processes the [30], [25], [19], [38], [39], [40], [22], [23], [24]). Considering features and if necessary sends a command to the pump (via the heterogeneous nature of PD and symptomatic profiles, this the mobile phone) in order to adjust the drug dose.
is not sufficient. Few publications focus on the recognition of REMPARK [7] uses a very similar setup, but takes PD multiple motor symptoms (e.g. [31], [18], [21], [26]), but even treatment a step further. Medication is not only dynamically those rarely consider enough symptoms for use in real-world administered, but freezing of gait (FOG) (episodes where a scenarios. In reality, multiple symptoms may overlap, thus patient simply freezes fo a few seconds) is actively tried to increasing the chance of false negatives and false positives.
overcome. The project itself is another EU research project For publications that focus on a single symptom, it is not clear with 11 partners in 7 countries. It started in October 2011 and how the occurrence of multiple symptoms affects the results is planed to end after 42 months in 2015. The consortium com- bines researchers, medical experts and industrialist partners.
So far no successful system for monitoring PD has been Form the hardware point of view, a body sensor and actu- build or is in wide use. This is mainly due to a lack of objective ator network is also employed. However two sensor platforms and reliable monitoring options (i.e. measurement of dopamine are utilized in this project (one worn around the wrist and levels). Furthermore, the heterogeneous nature of PD and the other on the waist). In addition to medication delivery symptomatic fluctuations thwart the building of such systems.
components, a functional electrical stimulation (FES) and / or Nonetheless, some projects pursue this very goal. HELP [6] auditive queuing device are going to be utilized to counteract and REMPARK [7] are two European research projects with a focus on providing a system for automatic recognition ofmotor symptoms and adjustment of medication administration in accordance with the patient’s ON / OFF state. Their basicidea is to reduce the time a patient spends in OFF state by As mentioned above, due to the lack of a practical option optimizing drug administration. Instead of manual adjustments to measure dopamine levels within the substantia nigra (basal by neurologists and pulse-based medication intake (i.e. in ganglia circuit), ON / OFF state estimation is based on the form of a pill taken every few hours), these projects thrive presence of motor symptoms. The difficulty of not being able for an automated, but dynamic, adjustment of a continues to directly infer drug needs makes the development of PD drug stimulation (i.e. in form of a pump). It is thought that monitoring systems (e.g. HELP [6], REMPARK [7]) unnec- a continues medication administration (at a low dose) can essarily hard. In addition, no unified JAVA-based framework reduce the overall drug consumption with equal or improved for time series analysis and data mining applications has been effectiveness when compared to pills (with a higher dose).
found which further increases development efforts. Thus, the Under this assumption, the complications and side-effects proposed outcomes of this research are: to compensate the lack would be decreased as well and enable a longer treatment with of a unified framework for time series analysis and provide algorithms for recognizing various motor symptoms of PD.
HELP [6] is 36 month European Union (EU) research To illustrate the scope and focus of this research, the project with 9 participants from Spain, Israel, Italy and Ger- pursued research questions are highlighted.
many. The consortium is a combination of medical profes-sionals, industry and research partners. The primary focus How can time series be represented? What is an of HELP is to build a closed-loop system for people with adequate architecture for a data mining and time Parkinson’s. Similar to a diabetes system where medication is administered with respect to measured glucose levels. However Depending on the data, time series can be represented in the HELP system, one of three distinct medication dosages in different ways. However, a unified representation is (i.e. low, high, bolus) is automatically administered based on a essential for a proper time series analysis framework set of features (e.g. amount of movement, location, etc.) which as it greatly simplifies implementation and architec- correlate with the patient’s ON / OFF state. A web-portal tural work. Analysis of time series can be tedious enables medical personal to view and override the decisions work, especially when algorithms need to be (re-) made by the HELP system. Here it is also possible to adjust the implemented from scratch or adapted every time a actual medication dosages associated with ”low”, ”high” and new data source (e.g. sensor) is added. Once com- ”bolus”. Furthermore the patient’s therapy, alerts, appointments pleted, such a framework enables the development can be viewed and modified if necessary [41].
and evaluation of not just PD indication algorithmsbut also has applications in other domains where The main components of the HELP system are: a body stream-based analysis is of relevance (e.g. financial sensor and actuator network (BS&AN), a service and network infrastructure and a web-portal. On the hardware level the Which Parkinson’s Disease symptoms can be rec- BS&AN consists of a sensor platform (with accelerometers, magnetic field sensors, gyroscopes and a temperature sensor), a Considering the wide range of symptoms (motor and mobile phone and either a subcutaneous pump (for use in more non-motor) shown by people with Parkinson’s, those advanced stages of the disease) or an intra-oral device (for use symptoms need to be identified that are likely to be in an early stage). The sensor platform is worn around the recognizable with off-the-shelf sensors. Whereas a waist and generates features regarding the presence of motor reasonable low interaction and unobtrusiveness for symptoms. These features are transmitted via ZigBee to the use in daily life are highly desirable. Among the mobile phone where they are forwarded to the main HELP innumerable features of sensor signals and algorithms that can be used to recognize symptoms, represen- expected to provide further insides on the matter in question tative features and appropriate algorithms are to be and could point out possible pitfalls.
A major part of future work focuses on data acquisition. An How can published state of the art techniquesfor recognizing motor symptoms of Parkinson’s effective evaluation of algorithms for indicating the presence of motor symptoms requires labeled data sets from varyingsensors, positions and sampling rates. It is planed to setup A large number of publications [37], [30], [25], and share a database (DB) with annotated and anonymous [19], [38], [39], [40], [22], [23], [24] related to data sets of patients with PD. It is the intention to provide PD symptoms indication only consider individual unrestricted access to the DB and thus aid the development of symptoms instead of multiple ones. They claim to more sophisticated algorithms of researchers around the planet.
have achieved high sensitivity and specificity, but itis unclear how the occurrence of multiple symptoms Preliminary research has revealed a great number of PD affects these results. The state of the art techniques related motor symptom indicators. However only a few of in this domain need to be reviewed. The applied AI them consider multiple symptoms as opposed to a single one.
algorithms, extracted features, utilized sensors and Signs and symptoms may overlap in PD, thus the presence of their positioning on the body should be identified.
multiple symptoms should be properly handled by indication How well do the new / improved approaches algorithms. Algorithms of this category need to be identified perform when compared to state of the art tech- and included in the evaluation process, requiring a detailed Competitive algorithms need to be compared to oneanother. Where possible state of the art algorithms, The recent discovery of a stream-based framework called their improved versions and newly developed ap- massive online analysis (MOA) [44] has raised the question of proaches should be evaluated (i.e. in terms of speed, whether to continue building a separate framework or support number of false positives and false negatives, online the development of an existing solution. The differences in the proposed framework of this research and existing MOAframework need to be identified and weighed against each These questions may be seen as a guideline of what is going to be done as part of this research effort. The primary focus ison improving and developing PD motor symptom indication algorithms. Consequently most effort and attention is devotedto answering the related research questions and less effort is PD is a serious neurodegenerative disorder of the central nervous system, which results in an increased rigidity andslowness. Additionally, the illness cannot be cured and patients have to rely on a proper treatment in order to avoid OFF Despite the early stage, research on PD and temporal data mining has been done. Furthermore, a software framework for Four research questions have been proposed. A question implementing and evaluating symptom indication algorithms is dealing with the absence of a unified and JAVA-based has been prepared. Most research on PD has already been time series framework. The remainder focuses on algorithms digested and compiled into a readable and structured docu- for indicating the presence of PD motor symptoms. The ment. As for the software framework, fundamental parts have research has only recently begun and much work remains to be been developed, implemented and tested. A number of well- done. Nonetheless, fundamental information has already been known design patterns [42] were employed to ease further development efforts and maintenance.
At the current state, further common temporal data mining approaches, including general and domain specific approaches, A. Samii, J. G. Nutt, and B. R. Ransom, “Parkinson’s disease,” The are being researched and compiled. In parallel, these ap- Lancet, vol. 363, no. 9423, pp. 1783 – 1793, 2004. [Online]. Available:http://www.sciencedirect.com/science/article/pii/S0140673604163058 proaches are implemented in the proposed framework.
http://jnnp.bmj.com/content/79/4/368.abstract Even though general information regarding PD’s pathology P. Andlin-Sobocki, B. J¨onsson, H.-U. Wittchen, and J. Olesen, and symptoms has been acquired, details on how motor symp- “Cost of disorders of the brain in Europe.” European Journal toms affect time series data is still rather unclear. Suitable of Neurology, vol. 12 Suppl 1, Jun. 2005. [Online]. Available: sensors and locations of such need to be identified. E.g.
http://dx.doi.org/10.1111/j.1468-1331.2005.01202.x in the case of tremor, the authors have seen the use of A. D. Korczyn, “Parkinson’s disease,” in International Encyclopedia accelerometers [25], [18], [21], gyroscopes [31], magnetic field Oxford: Academic Press, 2008, pp. 10 – 17. [Online]. Available: sensors [37] or electromyogram (EMG)-based signals [43], http://www.sciencedirect.com/science/article/pii/B9780123739605000289 [18], [21] at varying body locations. The optimal position and E. Tolosa, G. Wenning, and W. Poewe, “The diagnosis of parkinson’s sensor for motor symptoms of interest need to be identified.
disease.” Lancet Neurol, vol. 5, no. 1, pp. 75–86, 2006.
Complementary research on symptom indication algorithms in HELP, “Home-based empowered living for parkinson’s disease pa- the literature is planed and being undertaken. The results are tients,” jun 2012, www.help-aal.com.
REMPARK, “Personal health device for the remote and autonomous Network Applications in Electrical Engineering (NEUREL), 2010 10th management of parkinson’s disease,” jun 2012, www.rempark.eu.
Symposium on, sept. 2010, pp. 3 –6.
B. T. Cole, S. H. Roy, and S. H. Nawab, “Detecting freezing-of- Journal of neuropsychiatry and clinical neurosciences, vol. 14, gait during unscripted and unconstrained activity,” in Engineering no. 2, pp. 223–236; discussion 222, 2002. [Online]. Available: in Medicine and Biology Society,EMBC, 2011 Annual International http://www.ncbi.nlm.nih.gov/pubmed/11983801 Conference of the IEEE, 30 2011-sept. 3 2011, pp. 5649 –5652.
J. Cancela, M. Pansera, M. T. Arredondo, J. J. Estrada, M. Pastorino, L. Pastor-Sanz, and J. L. Villalar, “A comprehensive motor symp- tom monitoring and management system: The bradykinesia case,” in http://www.sciencedirect.com/science/article/pii/S0896627303005683 Engineering in Medicine and Biology Society (EMBC), 2010 Annual International Conference of the IEEE, 31 2010-sept. 4 2010, pp. 1008 G. Rigas, A. Tzallas, M. Tsipouras, P. Bougia, E. Tripoliti, D. Baga, http://www.sciencedirect.com/science/article/pii/S1873959808700243 D. Fotiadis, S. Tsouli, and S. Konitsiotis, “Assessment of tremor activity R. A. Armstrong, “Visual signs and symptoms of parkinson’s in the parkinson’s disease using a set of wearable sensors,” Information disease,” Clinical and Experimental Optometry, vol. 91, no. 2, pp.
Technology in Biomedicine, IEEE Transactions on, vol. PP, no. 99, p. 1, 129–138, 2008. [Online]. Available: http://dx.doi.org/10.1111/j.1444- S. Patel, K. Lorincz, R. Hughes, N. Huggins, J. Growdon, D. Standaert, C. A. Davie, “A review of parkinson’s disease,” British Medical M. Akay, J. Dy, M. Welsh, and P. Bonato, “Monitoring motor fluc- Bulletin, vol. 86, no. 1, pp. 109–127, 2008. [Online]. Available: tuations in patients with parkinson’s disease using wearable sensors,” http://bmb.oxfordjournals.org/content/86/1/109.abstract Information Technology in Biomedicine, IEEE Transactions on, vol. 13,no. 6, pp. 864 –873, nov. 2009.
A. Gustavsson, M. Svensson, F. Jacobi, C. Allgulander, J. Alonso,E. Beghi, R. Dodel, M. Ekman, C. Faravelli, L. Fratiglioni, L. J. Myers and C. D. MacKinnon, “Quanti cation of movement regular- B. Gannon, D. H. Jones, P. Jennum, A. Jordanova, L. J¨onsson, ity during internally generated and externally cued repetitive movements K. Karampampa, M. Knapp, G. Kobelt, T. Kurth, R. Lieb, in patients with parkinson’s disease.” in Neural Engineering, 2005.
M. Linde, C. Ljungcrantz, A. Maercker, B. Melin, M. Moscarelli, Conference Proceedings. 2nd International IEEE EMBS Conference on, A. Musayev, F. Norwood, M. Preisig, M. Pugliatti, J. Rehm, L. Salvador-Carulla, B. Schlehofer, R. Simon, H.-C. Steinhausen, T. Sugiura, N. Sugiura, K. Sugiyama, and T. Yokoyama, “Chaotic L. J. Stovner, J.-M. Vallat, P. V. den Bergh, J. van Os, P. Vos, approach to the quantitative analysis of parkinson’s disease,” in En- W. Xu, H.-U. Wittchen, B. J¨onsson, and J. Olesen, “Cost of disorders gineering in Medicine and Biology Society, 1998. Proceedings of the of the brain in europe 2010,” European Neuropsychopharmacology, 20th Annual International Conference of the IEEE, vol. 3, oct-1 nov vol. 21, no. 10, pp. 718 – 779, 2011. [Online]. Available: http://www.sciencedirect.com/science/article/pii/S0924977X1100215X S. L. Smith and K. Shannon, “Vector-based analysis of motor activities C. Mathers, D. M. Fat, J. T. Boerma, and W. H. Organization, in patients with parkinson’s disease,” in EUROMICRO 97. ’New Fron- The global burden of disease : 2004 update.
tiers of Information Technology’. Short Contributions., Proceedings of Organization, Geneva, Switzerland :, 2008. [Online]. Available: the 23rd Euromicro Conference, sep 1997, pp. 50 –55.
http://www.who.int/healthinfo/global burden disease/GBD report O. ˇSprdl´ık, Z. Hur´ak, M. Hoskovcov´a, and E. R˚uˇziˇcka, “Tremor analysis by decomposition of acceleration into gravity and inertial acceleration D. J. Gelb, E. Oliver, and S. Gilman, “Diagnostic criteria for parkinson using inertial measurement unit,” in Information Technology and Appli- disease,” Arch Neurol, vol. 56, no. 1, pp. 33–39, 1999. [Online].
cations in Biomedicine, 2009. ITAB 2009. 9th International Conference Available: http://archneur.ama-assn.org/cgi/content/abstract/56/1/33 A. Salarian, H. Russmann, C. Wider, P. R. Burkhard, F. J. G. Vinger- hoets, and K. Aminian, “Quantification of tremor and bradykinesia in parkinson’s disease using a novel ambulatory monitoring system,” http://www.sciencedirect.com/science/article/pii/S1353802001000244 Biomedical Engineering, IEEE Transactions on, vol. 54, no. 2, pp. 313 N. L. W. Keijsers, M. W. I. M. Horstink, and S. C. A. M. Gielen, “Automatic assessment of levodopa-induced dyskinesias in daily life K. Niazmand, A. Kalaras, H. Dai, and T. C. Lueth, “Comparison of by neural networks,” Movement Disorders, vol. 18, no. 1, pp. 70–80, methods for tremor frequency analysis for patients with parkinson’s 2003. [Online]. Available: http://dx.doi.org/10.1002/mds.10310 disease,” in Biomedical Engineering and Informatics (BMEI), 2011 4th B. T. Cole, S. H. Roy, C. J. De Luca, and S. H. Nawab, “Dynamic neural International Conference on, vol. 2, oct. 2011, pp. 693 –697.
network detection of tremor and dyskinesia from wearable sensor data,” K. Niazmand, K. Tonn, Y. Zhao, U. M. Fietzek, F. Schroeteler, in Engineering in Medicine and Biology Society (EMBC), 2010 Annual K. Ziegler, A. O. Ceballos-Baumann, and T. C. Lueth, “Freezing of International Conference of the IEEE, 31 2010-sept. 4 2010, pp. 6062 gait detection in parkinson’s disease using accelerometer based smart clothes,” in Biomedical Circuits and Systems Conference (BioCAS), R. Arvind, B. Karthik, N. Sriraam, and J. K. Kannan, “Automated 2011 IEEE, nov. 2011, pp. 201 –204.
detection of pd resting tremor using psd with recurrent neural network E. Jovanov, E. Wang, L. Verhagen, M. Fredrickson, and R. Fratangelo, classifier,” in Advances in Recent Technologies in Communication and “defog - a real time system for detection and unfreezing of gait of Computing (ARTCom), 2010 International Conference on, oct. 2010, parkinson’s patients,” in Engineering in Medicine and Biology Society, 2009. EMBC 2009. Annual International Conference of the IEEE, sept.
E. Bakstein, K. Warwick, J. Burgess, O. Stavdahl, and T. Aziz, “Features for detection of parkinson’s disease tremor from local field potentials M. B¨achlin, M. Plotnik, D. Roggen, I. Maidan, J. M. Hausdorff, of the subthalamic nucleus,” in Cybernetic Intelligent Systems (CIS), N. Giladi, and G. Troster, “Wearable assistant for parkinson’s disease 2010 IEEE 9th International Conference on, sept. 2010, pp. 1 –6.
patients with the freezing of gait symptom,” Information Technology S. H. Roy, B. T. Cole, L. D. Gilmore, C. J. De Luca, and S. H.
in Biomedicine, IEEE Transactions on, vol. 14, no. 2, pp. 436 –446, Nawab, “Resolving signal complexities for ambulatory monitoring of motor function in parkinson’s disease,” in Engineering in Medicine and M. B¨achlin, D. Roggen, G. Troster, M. Plotnik, N. Inbar, I. Meidan, Biology Society,EMBC, 2011 Annual International Conference of the T. Herman, M. Brozgol, E. Shaviv, N. Giladi, and J. M. Hausdorff, IEEE, 30 2011-sept. 3 2011, pp. 4832 –4835.
“Potentials of enhanced context awareness in wearable assistants for M. Djuri´c-Joviˇci´c, N. S. Joviˇci´c, I. Milovanovi´c, S. Radovanovi´c, parkinson’s disease patients with the freezing of gait syndrome,” in N. Kresojevi´c, and M. B. Popovi´c, “Classification of walking patterns Wearable Computers, 2009. ISWC ’09. International Symposium on, in parkinson’s disease patients based on inertial sensor data,” in Neural Y. Su, C. R. Allen, D. Geng, D. Burn, U. Brechany, G. D. Bell, andR. Rowland, “3-d motion system (”data-gloves”): application for parkin-son’s disease,” Instrumentation and Measurement, IEEE Transactionson, vol. 52, no. 3, pp. 662 – 674, june 2003.
A. P. L. B´o and, P. Poignet, and C. Geny, “Pathological tremor and vol-untary motion modeling and online estimation for active compensation,”Neural Systems and Rehabilitation Engineering, IEEE Transactions on,vol. 19, no. 2, pp. 177 –185, april 2011.
K. Niazmand, K. Tonn, A. Kalaras, U. M. Fietzek, J. H. Mehrkens, andT. C. Lueth, “Quantitative evaluation of parkinson’s disease using sensorbased smart glove,” in Computer-Based Medical Systems (CBMS), 201124th International Symposium on, june 2011, pp. 1 –8.
C. D. Marchis, M. Schmid, and S. Conforto, “An optimized method fortremor detection and temporal tracking through repeated second ordermoment calculations on the surface emg signal,” Medical Engineering& Physics, vol. 34, no. 9, pp. 1268 – 1277, 2012. [Online]. Available:http://www.sciencedirect.com/science/article/pii/S1350453311003444 C. Ahlrichs, A. Sam`a, J. R. Simon, S. Herrlich, and A. Rodr´ıguez-Molinero, “Help: Optimizing treatment of parkinson’s disease patients,”in 3rd International Conference on the Elderly and New Technologies,Castell´on, Spain, Apr. 2012.
E. Gamma, R. Helm, R. Johnson, and J. M. Vlissides, Design Patterns:Elements of Reusable Object-Oriented Software.
J. L. Dideriksen, F. Gianfelici, L. Z. P. Maneski, and D. Farina, “Emg-based characterization of pathological tremor using the iterated hilberttransform,” Biomedical Engineering, IEEE Transactions on, vol. 58,no. 10, pp. 2911 –2921, oct. 2011.
MOA, “Massive online analysis,” jul 2012, http://moa.cs.waikato.ac.nz.

Source: http://www.comnets.uni-bremen.de/cewit-tzi-workshop-2013/PDF/papers/Ahlrichs.pdf