The localization of human brain activity, i.e., neuroelectromagnetic source estimation, is an inversion problem where we have a limited amount of electromagnetic data measured from outside the head and a much larger number of neuron populations and their possible configurations, which can produce similar-looking measurement curves. In order to limit the possible electromagnetic source configurations to only one, additional assumptions about the nature of the activity are needed.
In his research, Lahtinen focused on localising focal brain activity, which is limited to a tiny area. Examples of these localised activities include the focus of activity in focal epilepsy, and the stimulation signal from the median nerve. From the latter, we know that the activity begins in the brain's deep structures, the brainstem and the thalamus, and after that moves to the primary somatosensory cortex.
"Somatosensory activity has been studied as being limited to a small area already in the 1980s, but there have been surprisingly few attempts to develop a method to estimate the source specifically for focal activity," he says.
Focal epilepsy can be localised with clinical-level accuracy using the standardisation technique
Recently, interest in source estimation to support clinical work has increased, and this has also created new challenges. For example, in several pathological cases, knowledge of brain activity in the deep parts of the brain would be particularly important.
Previous studies have shown that a method called standardized low-resolution brain electromagnetic tomography can be accurate, even when dealing with deep activity. But it is not always reliable: When it works, the method is excellent, but when it goes wrong, it is by no means accurate.
Lahtinen thinks that if we knew about these situations where errors could happen, we could avoid them. In his article, he presented the reasons for the accuracy of the technique and the conditions under which it might fail.
"Understanding standardisation technology in a new way opened the doors to two new methods based on the same technology," he says.
One of the standardised methods presented by Lahtinen is tailored for focal estimation, and the other is intended to evaluate how brain activity changes over time during the measurement period. The methods are designed for clinical use, requiring measurement noise tolerance and excellent positioning accuracy, both in the surface and deep layers of the brain.
Public defence on Friday 7 March
The doctoral dissertation of MSc Joonas Lahtinen in the field of mathematics titled Mathematical Methods for the Neuroelectromagnetic Inverse Problem for Focal and Unbiased Brain Imaging will be publicly examined at the Faculty of Engineering and Natural Sciences at Tampere University at 12 o’clock on Friday 7 March 2025 at Hervanta campus, in Tietotalo building at auditorium TB109 (Korkeakoulunkatu 1, Tampere).
The Opponent will be Professor Matti Hämäläinen from Aalto University. The Custos will be Professor Sampsa Pursiainen from the Faculty of Information Technology and Communication at the University of Tampere.
The doctoral dissertation is available online.
The public defence can be followed via remote connection.