Skip to content Search

Mental health diagnosis

Most mental disorders are highly complex and have high phenotypic variability, partially vague diagnostic criteria, and a significant overlap ratio. A single clinical examination is not representative and is susceptible to the framing effect and other cognitive biases. The patient’s subjective experience is often precious but also severely neglected due to its vagueness, variability and an idiosyncratic presentation of symptoms. The overwhelming majority of psychiatric disorders appear to be dimensional in nature. Consequently, their categorizations become artificial and debatable.

Mental health diagnosis remains prone to systemic and predictable cognitive errors resulting from not-fully-conscious simplistic inference schemes called heuristics, which may lead to epistemic injustice affecting medical assessments.

To reduce such epistemic injustice, we aim to develop tools for computer-assisted recognition of experience patterns in selected mental conditions by incorporating a semantic network of lived experience founded upon a hybrid database combining both third-person medical data and deep phenomenological first-person reports. This is also important since many mental conditions are increasingly considered not mental disorders per se but neurodiverse conditions (which means they entail only statistically atypical, yet often hyper- or hypo-functional perceptions and behaviors).

The elements most often overlooked in the development of such tools are: a) the first-person viewpoint of patients, which is usually not amenable to quantification and quantitative analysis, not only due to its nature but also the lack of a sufficiently extensive database of patient self-reports that could be subjected to appropriate analysis; b) objective, quantitative data from external sources (e.g. social media profiles), which are currently largely neglected in the diagnostic process.

Schizophrenia and autism

The diagnosis of schizophrenia is a long process subjected to multiple biases. Often guided by inexplicable intuition, it may be assigned when a clinical presentation does not match strict diagnostic criteria. There are several excellent predictors of schizophrenia, which are generally hard to take into account in a typical clinical setting through operationalized diagnosis, but may be detectable by preconscious intuitions and augmented intelligence tools mimicking it. Such automated tools facilitating the diagnostic process could limit the risk of biases

The diagnosis of autism is extremely time-consuming and requires highly trained professionals to ensure its accuracy. In addition, traditional methods of combating prejudice against people with autism are ineffective and the social costs are usually enormous if only in the form of wasting the potential of people with autism (research shows that many people on the autism spectrum are characterized by a ‘hyper-systematizing’ type of cognitive skills, which predisposes them naturally to IT-related employment). The main obstacles to fully realizing and exploiting this potential are late diagnostic recognitions affecting social functioning and low public awareness of the specific functioning of people on the autism spectrum.

Semantic networks

Rozwiązaniem jest obliczeniowy model danych mikrofenomenalnych umożliwiający zautomatyzowaną analizę doświadczeń osób w spektrum schizofrenii i autyzmu, a więc obejmujący istotny, choć często ignorowany aspekt diagnozy oraz zmierzający do przezwyciężenia uprzedzeń społecznych wobec tych zaburzeń. Algorytmy NLP mogą analizować wzorce i cechy języka w korpusach odkrywając znaczące powiązania dot. zdrowia psychicznego. Można je np. wykorzystać do analizy wyrażonego w tekście sentymentu i tonu emocjonalnego, co pozwala zidentyfikować wzorce wyrażające niepokój, depresję, lęk itp. lub konkretne pojęcia reprezentujące symptomy.

Algorytmy modelowania tematycznego mogą grupować powiązane dokumenty lub zdania, co może pomóc w identyfikacji traumy lub stresu; zmiany w użyciu zaimków (np. przejście z „ja” na „ty”) lub obecność określonych znaczników językowych może wskazywać na zmiany w postrzeganiu siebie; analiza behawioralna za pomocą tekstu może pomóc w identyfikacji wycofania społecznego, izolacji lub uprzedzeń poznawczych.  Zaletą sieci semantycznych jest to, że można je wykorzystać do reprezentowania znaczeń (w tym metafor) języka naturalnego w sposób łatwy do interpretacji zarówno przez sztuczną inteligencję, jak i ludzi, co może ułatwić przezwyciężenie nieufności do narzędzi diagnostycznych wspomaganych komputerowo. Nasze podejście może potencjalnie wyeliminować niektóre uprzedzenia diagnostyczne oraz uwzględnić głębsze rozumienie istoty doświadczenia autystycznego i schizofrenicznego.

Research team leader


Marcin Moskalewicz

Marcin Moskalewicz is Associate Professor at Poznan University of Medical Sciences, where he is the head of Philosophy of Mental Health Unit, and Associate Professor at the Institute of Philosophy, Marie Curie-Sklodowska University in Lublin; convener of the Phenomenology and Mental Health Network at The Collaborating Centre for Values-based Practice in Health and Social Care, St. Catherine’s College, Oxford; Executive Committee Member at the Royal College of Psychiatrists Philosophy Special Interest Group as well as Board Member of Open Seminars on Philosophy and Psychiatry Foundation in Warsaw.

Moskalewicz studied at Adam Mickiewicz University in Poznan, University of California at Berkeley, and Rijksuniversiteit Groningen; PhD 2009 (summa cum laude), habilitation 2018 (Polish Academy of Sciences). Moskalewicz specializes in transdisciplinary research at the intersection of health sciences, psychiatry and phenomenology, especially diagnostic expertise, issues of neurodiversity and application of computational psychiatry tools to the study of lived experience. He published over 50 scientific papers.

Prime Minister of Poland Prize (2010), Foundation for Polish Science “Start” Scholarship (2011) and “Monografie” book prize (2013), Polish Ministry of Science Outstanding Young Scientists Scholarship (2015-2018), Individual Scientific Award of PUMS Rector (2013 and 2022); PI in National Science Center grants Sonata (2011-2014) and Sonata Bis (2022-), and Polish-German Science Foundation (2015) and Polish Ligue Against Cancer (2019-2020) grants.

International cooperation: Marie Curie Fellow at Rijksuniversiteit Groningen (2005 and 2007), EURIAS Fellow at ETH Zurich (2015), Senior Fulbright Scholar at Texas A&M University (2016), Marie Curie Fellow at the Faculty of Philosophy, University of Oxford and Senior Member of Christ Church College (2016-2017), Fellow at TORCH, The Oxford Research Centre in the Humanities (2017-2019), Humboldt Research Fellow at Psychiatric Clinic, University of Heidelberg (2022).

Other research groups and teams

  • Sustainable Machine Learning For Autonomous Machines Our solutions could potentially be used in drones as a tool supporting the protection of national parks, including animals against poaching. They allow for fast and efficient monitoring of large land areas in remote locations...
    Bartosz Zieliński
  • Continual Machine Learning Artificial neural networks are powerful models used in various fields because they can learn to represent useful features. When we train neural networks, we usually start with random settings and adjust them based on the data available at that moment. However, this approach is different from how humans learn. Humans continuously build on their knowledge over time; they are lifelong learners.
    Bartłomiej Twardowski
  • Precision forestry The use of remote sensing data in obtaining information about forests has a long, almost 100-year history.
    Krzysztof Stereńczak