PSYKOSE: A Motor Activity Database of Patients with Schizophrenia

Abstract

Using sensor data from devices such as smart-watches or mobile phones is very popular in both computer science and medical research. Such movement data can predict certain health states or performance outcomes.However, in order to increase reliability and replication of the research it is important to share data and results openly. In medicine, this is often difficult due to legal restrictions or to the fact that data collected from clinical trials is seen as very valuable and something that should be kept "in-house". In this paper, we therefore present PSYKOSE, a publicly shared dataset consisting of motor activity data collected from body sensors. The dataset contains data collected from patients with schizophrenia. Schizophrenia is a severe mental disorder characterized by psychotic symptoms like hallucinations and delusions, as well as symptoms of cognitive dysfunction and diminished motivation. In total, we have data from 22 patients with schizophrenia and 32 healthy control persons. For each person in the dataset, we provide sensor data collected over several days in a row. In addition to the sensor data, we also provide some demographic data and medical assessments during the observation period. The patients were assessed by medical experts from Haukeland University hospital. In addition to the data, we provide a baseline analysis and possible use-cases of the dataset.

Comparison of Manual and Automated Feature Engineering for Daily Activity Classification in Mental Disorder Diagnosis

“…The method presented in [14] uses just 3 features for depression classification: mean activity level, standard deviation, and proportion of zeros (fraction of minutes with zero activity). All works described above use the Depresjon dataset; for schizophrenia classification, we failed to find any paper apart from the Psykose dataset paper [2], in which baselines are presented, using mean, standard deviation, and proportion of zeros.…”

“…Multiple techniques are described in [25] and [26], including popular methods such as holdout and cross-validation, but also more sophisticated bootstrap-based algorithms such as 0.632 estimator. The problem of properly measuring classifier performance on mental disorder datasets, because of their small sample size, has been mentioned in [2], but only simple cross-validation is suggested.…”

“…Actigraph IoT devices, very similar to watches and often combined with watch functionality, can be used to record the continuous 1D signal with the wearer's activity, based on the accelerometers built into them. Data from such sensors has been applied in the mental health domain to support the diagnosis of bipolar disorder [11], ADHD [12], depression [1] and schizophrenia [2]. We focus on the last two since datasets for classification have been published for them.…”

“…While neural networks such as Gated Recurrent Unit (GRU) or Long-Short Term Memory (LSTM) can work with it, they are known to require a significant amount of data to train, and available datasets are very small. For those reasons, the approach taken by most papers is a classical 2-step process: extract statistical features and use a regular classifier for tabular data [2,13,15]. This approach can be performed manually, where researchers choose features themselves.…”

“…We use Depresjon [1] and Psykose [2] datasets, for diagnosis of depression and schizophrenia, respectively. In both cases, they contain 24 h motor activity signals measured with actigraph watches.…”