top of page


Publications in progress


  • Davies, A. L., Coolen, A. C. C., & Galla, T. (2021). Retarded kernels for longitudinal survival analysis and dynamic prediction(under review) arXiv:2110.11196 (Cornell University).

  • Jonker, M. A., Pazira, H., & Coolen, A. C. (2024). Bayesian Federated Inference for regression models with heterogeneous multi-center populations. (under review)

Other news

AISN Consortium Meeting

Saddle Point Science Europe, a technical partner in the AISN (Artificial Intelligence for Stroke Neurorehabilitation) consortium, works alongside esteemed companies, research institutes, and clinicians from across Europe. The consortium, dedicated to advancing stroke rehabilitation through Rehabilitation Gaming Systems (RGS), entrusted Saddle Point Science Europe with the development of mathematical prognostic and diagnostic models leveraging RGS parameters and data.

On February 14th and 15th in Interlaken, Switzerland, the AISN consortium meeting was held to gather experts to present their work on stroke rehabilitation. Over the course of two days, technical partners had the opportunity to show their contributions to the AISN pipeline, which involves stroke prognostics, diagnostics, and treatment recommendation systems.  

Theodore Nikoletopoulos discussed the details of a mathematical model predicting clinical scales (numbers that describe stroke recovery in terms of certain scales used by clinicians). This model is essential to the AISN pipeline's decision support system and could support  stroke rehabilitation by determining optimal treatment parameters. Theodore also introduced methods to infer clinical scales from RGS data developed at Saddle Point Science, offering an approach for diagnosis based on RGS sessions.

Saddle Point Science Europe is looking forward to ongoing collaboration and joining efforts from all stakeholders. We are proud to play a part in shaping the future of stroke neurorehabilitation.


I4MDS General meeting:
Big data in Cancer Research 

On February 9th and 10th the 2nd I4MDS General Meeting was held in The Hague, with support from the European Hematology Association. MDS, which stands for Myelodysplastic Neoplasm (see e.g. this recent review), is a disease of blood cells that is a specific precursor stage of acute myeloid leukemia.


The I4MDS consortium is an international team of medical researchers, mainly from Europe and the USA,  that specialize in the study of MDS and the immune system. They are together trying to identify signals in the blood cells produced by the human immune system with which to predict for individual patients how their disease is most likely to develop over time, and which immune therapy (if any) might work for them. I had the honour of being invited to speak at the above meeting, on the topic `Big data in cancer immunology – dangers and opportunities’.


Ton Coolen's talk focused on methodology for dealing with dimension mismatch (too many variables, too few samples), such as overfitting correction pipelines and Bayesian federated Inference, and for responder identification in clinical trials. These topics proved to be quite relevant to the I4MDS ambitions, and we can see many fruitful collaborations with the I4MDS team in the future!

bottom of page