% IMPORTANT: The following is UTF-8 encoded.  This means that in the presence
% of non-ASCII characters, it will not work with BibTeX 0.99 or older.
% Instead, you should use an up-to-date BibTeX implementation like “bibtex8” or
% “biber”.

@PHDTHESIS{SahandiFar:1014708,
      author       = {Sahandi Far, Mehran},
      title        = {{S}ensor-based {A}ssessments as a {D}isease {P}rogression
                      and {T}reatment {B}iomarker for {N}europsychiatric
                      {D}iseases},
      school       = {HHU Düsseldorf},
      type         = {Dissertation},
      reportid     = {FZJ-2023-03403},
      pages        = {157},
      year         = {2023},
      note         = {Dissertation, HHU Düsseldorf, 2023},
      abstract     = {The growing trend of personalised health care and remote
                      monitoring has led to increased interest in using embedded
                      sensors in portable smart devices (smartphones and
                      smartwatches) in clinical studies. Health-related data
                      collected from such devices are
                      referredtoasDigitalBiomarkers(DBs).
                      Unliketraditionalin-clinicassessmentmethods, DBs provide
                      cost-effective, objective, and ecologically valid data. DBs
                      enable clinical studies to recruit a larger and more diverse
                      population. Furthermore, DBs provide high temporal and
                      spatial resolution data, which increase the chance of
                      gaining a comprehensive understanding of disease
                      progression.Neurodegenerative diseases, due to their lack of
                      accessible and objective assessment tools, have been a
                      primary focus for the DBs research community. Parkinson's
                      disease (PD) is particularly well-suited for studying DBs
                      due to its heterogeneous onset age, symptom prevalence,
                      severity progression rate, and multiple aspects of the
                      disease. Therefore, there is a need to integrate DBs and
                      remote assessment into the routine clinical evaluation of
                      PD. However, using DBs for PD in non-controlled, at-home
                      settings poses practical challenges that have hindered this
                      goal. Additionally, the longitudinal stability of DBs
                      collected in such settings has not yet been thoroughly
                      investigated, with previous studies limited to in-lab
                      settings. Thus, this thesis aims to provide insight into how
                      remote monitoring in an at-home environment alongside the
                      data collection methods can be leveraged to improve the way
                      PD is assessed.The first section of this dissertation
                      focuses on introducing a platform named "JTrack", designed
                      for remote disease monitoring and to address technical
                      aspects such as security, privacy, modularity, and
                      reusability. This platform aims to provide a comprehensive
                      solution for clinical studies involving multiple aspects of
                      various diseases. In addition, this section assesses the
                      agreement between features collected through "JTrack" with
                      two widely used stationary systems for analysing gait and
                      balance, demonstrating the potential of using smartphones
                      and particularly the "JTrack" platform in future clinical
                      studies. The second part of this thesis investigates the
                      potential of using various commonly reported features in PD
                      studies as biomarkers. To do this, we first investigate
                      these features' test-retest reliability and longitudinal
                      stability, considering how the timescale may affect their
                      stability. Next, we use various machine learning algorithms
                      to assess the ability of these features to differentiate
                      between PD and HC. Also, we evaluated the influence of
                      different confounding factors such as comorbidities, age,
                      and sex on the prediction performance of the machine
                      learning algorithms. For this, the various tasks (gait,
                      balance, voice, and tapping) of the m-Power database,
                      collected remotely and in a self-managed setting, were
                      investigated.Overall, this thesis discusses the potential
                      and limitations of using smartphones for remote assessment
                      of PD. It examines the possible sources of confounding
                      factors related to DBs in remote and self-managed collection
                      methods. It also highlights the need to develop more
                      controlled, standardised, sensitive, and reliable DBs before
                      taking them into any clinical application. This thesis also
                      introduces a new DBs platform for remote assessment, which
                      can be leveraged for various types of disease.},
      cin          = {INM-7},
      cid          = {I:(DE-Juel1)INM-7-20090406},
      pnm          = {5252 - Brain Dysfunction and Plasticity (POF4-525)},
      pid          = {G:(DE-HGF)POF4-5252},
      typ          = {PUB:(DE-HGF)11},
      doi          = {10.34734/FZJ-2023-03403},
      url          = {https://juser.fz-juelich.de/record/1014708},
}