Skip to main content

Refine your search

Elderly patient performing a drawing test.

Doctoral defence of Vandad Imani, MSc, 24 May 2024: New machine learning applications combine multimodal data for early diagnosis of brain disorders

The doctoral dissertation in the field of Artificial Intelligence will be examined at the Faculty of Health Sciences at Kuopio campus. 

What is the topic of your doctoral research? Why is it important to study the topic?

The topic of my doctoral research is the development of novel machine learning (ML) applications for early diagnosis of brain disorders, particularly Alzheimer's disease and cognitive decline, using multimodal imaging data alongside cognitive, behavioral, and genetic data.

Studying this topic is important due to increasing complexity of brain disorders and the need for more accurate diagnostic approaches. Early detection of brain disorders allows for timely intervention and management, potentially improving treatment outcomes and quality of life for affected individuals.

Moreover, the integration of multimodal imaging data, cognitive tests, and genetic information presents a holistic approach to understanding brain health and disease progression. By developing robust computational methods to analyse diverse data sources, my research aims to enhance our capabilities in early detection and understanding of brain abnormalities, ultimately contributing to advancements in medical diagnostics and patient care.

What are the key findings or observations of your doctoral research?

The key findings of this doctoral research focus on advancing machine learning (ML) applications for the diagnosis of brain disorders through multimodal neuroimaging data. The thesis introduces comprehensive ML methodologies tailored specifically to address the inherent challenges of multimodal data, such as data heterogeneity, high dimensionality, and the presence of noisy or incomplete data. These technical advancements significantly enhance the diagnostic capabilities for early detection of disorders like Alzheimer's.

A notable contribution of this research is the development of novel supervised learning algorithms. These algorithms are optimized to leverage multiple data sources—including MRI, PET, and cognitive tests—effectively integrating these to improve early diagnosis of complex brain disorders. This approach not only elevates the precision but also the reliability of the diagnoses.

Another major methodological advancement is the implementation of sophisticated data imputation techniques. These techniques are crucial for managing missing data in longitudinal studies, a common obstacle in neuroimaging research. By ensuring robust data integration over time and across different modalities, the research supports more consistent and reliable analysis outcomes.

Furthermore, the thesis presents innovative feature selection methods designed to address the high dimensionality of multimodal data. These methods enhance the performance of ML models by identifying and prioritizing the most relevant features, thereby improving the accuracy and reliability of the diagnostic models.

Overall, these methodological contributions represent significant advancements in the use of ML for neuroimaging and set a foundation for further research in the diagnosis and understanding of brain disorders. The techniques developed are not only applicable to Alzheimer's disease and cognitive decline but also hold potential for broader applications in medical diagnostics.

How can the results of your doctoral research be utilised in practice?

The results of my doctoral research can be applied practically to facilitate early diagnosis of brain disorders, personalise treatment plans based on individual patient profiles, develop versatile biomarkers for disease monitoring and treatment response assessment, provide decision support tools for healthcare professionals, and advance research methodologies in medical diagnostics and predictive analytics.

What are the key research methods and materials used in your doctoral research?

The doctoral research process involved using multimodal imaging data such as MRI and PET scans, alongside cognitive and genetic data. Machine learning algorithms were applied to develop predictive models for early diagnosis of brain disorders. Innovative methodologies were employed to address challenges like data heterogeneity and high dimensionality, including data imputation techniques and feature selection methods. Statistical analysis validated the effectiveness of these approaches in enhancing diagnostic precision and understanding brain abnormalities.

The doctoral dissertation of Vandad Imani, MSc, entitled Machine Learning for Early Diagnosis of Brain Disorders Based on Multimodal Data will be examined at the Faculty of Health Sciences. The Opponent in the public examination will be Professor Miguel Bordallo Lopez of the University of Oulu and the Custos will be Professor Jussi Tohka of the University of Eastern Finland.

Doctoral defence 


Doctoral dissertation 

For further information, please contact:

Vandad Imani, MSc,, 0449516463