Lars Forsberg

FAQ About Lars Forsberg

• What is Lars Forsberg's address?

  Lars Forsberg is located at
  Stockholm, Stockholm, Sweden

• What is Lars Forsberg's job title?

  Lars Forsberg is a
  I am an engineering physicist with a PhD in medical science and 20 years of experience in analysing real-world data, including brain imaging of ageing and multiple sclerosis data. Brain Image Analysis My focus in brain image analysis has been on developing advanced tools and techniques that leverage machine learning to better understand brain morphology, including the following examples: Developing visualisation tools in C/C++/Objective-C using the GPU for rendering and computations. One notable tool facilitated the creation of a regional brain atlas and corresponding segmentation software (Forsberg et al., Medical Image Analysis, 2017). Developing tools in CUDA and OpenCL to perform massive structural brain correlation analyses on the GPU. This was used to identified hub-like features in subcortical regions of the brain (Forsberg et al., Neuroimage, 2019). Using neural networks (Variational Autoencoders) in Keras to further explored the latent space of structural brain images, providing new insights into complex neuroimaging data (Forsberg et al., Manuscript, to be published). Multiple Sclerosis My focus in multiple sclerosis research is on developing federated analysis techniques, particularly for scenarios where different sites cannot share data. One study using federated analysis examined the proportion of secondary progressive multiple sclerosis patients across five countries (Forsberg et al., Multiple Sclerosis Journal - Experimental, Translational and Clinical, 2023). This work has continued to involve federated learning. At ECTRIMS 2023, I presented a federated learning tool written in R that enabled logistic regression between two countries (Sweden and Denmark) as if the data were pooled. Unlike most federated learning solutions that require each registry to install a server, this solution only requires running R Studio at each registry. This work will be further developed to include Cox regression and propensity score weighting. Another key area of research is using Patient Reported Outcomes (PROs) as a proxy for clinical data. At ECTRIMS 2022, I presented findings demonstrating the link between short-term changes in clinical disability (using EDSS) and long-term changes in PRO scores (using MSIS-29), by applying both regression analyses and neural networks. I also demonstrated a high correlation between a neural network predictor of EDSS given the input of MSIS-29. Additionally, I have contributed to two post-authorisation safety studies (PASS) in multiple sclerosis, ensuring adherence to regulatory protocols and sponsor timelines. Karolinska Institutet

• What industry does Lars Forsberg currently work in?

  Lars Forsberg currently works in the
  Karolinska Institutet

• What is Lars Forsberg's area of expertise?

  Lars Forsberg specializes in
  I am an engineering physicist with a PhD in medical science and 20 years of experience in analysing real-world data, including brain imaging of ageing and multiple sclerosis data. Brain Image Analysis My focus in brain image analysis has been on developing advanced tools and techniques that leverage machine learning to better understand brain morphology, including the following examples: Developing visualisation tools in C/C++/Objective-C using the GPU for rendering and computations. One notable tool facilitated the creation of a regional brain atlas and corresponding segmentation software (Forsberg et al., Medical Image Analysis, 2017). Developing tools in CUDA and OpenCL to perform massive structural brain correlation analyses on the GPU. This was used to identified hub-like features in subcortical regions of the brain (Forsberg et al., Neuroimage, 2019). Using neural networks (Variational Autoencoders) in Keras to further explored the latent space of structural brain images, providing new insights into complex neuroimaging data (Forsberg et al., Manuscript, to be published). Multiple Sclerosis My focus in multiple sclerosis research is on developing federated analysis techniques, particularly for scenarios where different sites cannot share data. One study using federated analysis examined the proportion of secondary progressive multiple sclerosis patients across five countries (Forsberg et al., Multiple Sclerosis Journal - Experimental, Translational and Clinical, 2023). This work has continued to involve federated learning. At ECTRIMS 2023, I presented a federated learning tool written in R that enabled logistic regression between two countries (Sweden and Denmark) as if the data were pooled. Unlike most federated learning solutions that require each registry to install a server, this solution only requires running R Studio at each registry. This work will be further developed to include Cox regression and propensity score weighting. Another key area of research is using Patient Reported Outcomes (PROs) as a proxy for clinical data. At ECTRIMS 2022, I presented findings demonstrating the link between short-term changes in clinical disability (using EDSS) and long-term changes in PRO scores (using MSIS-29), by applying both regression analyses and neural networks. I also demonstrated a high correlation between a neural network predictor of EDSS given the input of MSIS-29. Additionally, I have contributed to two post-authorisation safety studies (PASS) in multiple sclerosis, ensuring adherence to regulatory protocols and sponsor timelines.

• Which languages does Lars Forsberg speak? 

  Lars Forsberg speaks the following languages:
  No languages available.

• Which industries has Lars Forsberg previously worked in?   

  Lars Forsberg has prior experience in the following industries:
  Karolinska Institutet , Cubist , Icelandic Heart Association , Raforninn , Pdc Center For High Performance Computing At Kth

• What is Lars Forsberg’s work experience?  

  Lars Forsberg has the following professional experience:
  Karolinska Institutet , Karolinska Institutet , Cubist , Icelandic Heart Association , Raforninn , Rapidsmountain , Pdc Center For High Performance Computing At Kth .

• How can I find Lars Forsberg on LinkedIn?  

  You can view Lars Forsberg’s
  LinkedIn profile at LinkedIn URL.

• What are Lars Forsberg’s work email addresses?  

  Lars Forsberg’s work email addresses are
  [email protected] .

• What company does Lars Forsberg work for?  

  Lars Forsberg works at
  Karolinska Institutet