Eneko Uruñuela
@enekourunuela.com
Postdoctoral Researcher at University of Calgary 🇨🇦
Deep Learning | Computer Vision | Medical Imaging
Passionate about med-tech, personal knowledge management, and personal growth 🌱
Deep Learning | Computer Vision | Medical Imaging
Passionate about med-tech, personal knowledge management, and personal growth 🌱
I'm on that same boat...
November 3, 2025 at 7:57 PM
I'm on that same boat...
Done! Let me know if you need more info or links to our code.
October 16, 2025 at 5:41 PM
Done! Let me know if you need more info or links to our code.
We are still seeing issues as of right now.
October 16, 2025 at 4:01 PM
We are still seeing issues as of right now.
My heart is not ready for a crazy game like that one 😅
May 7, 2025 at 12:02 AM
My heart is not ready for a crazy game like that one 😅
And complain less (I hope)
May 1, 2025 at 8:19 PM
And complain less (I hope)
True, but you would still need a ground truth to learn from, right?
I’m more than happy to discuss ideas in this topic.
I’m more than happy to discuss ideas in this topic.
April 25, 2025 at 4:52 PM
True, but you would still need a ground truth to learn from, right?
I’m more than happy to discuss ideas in this topic.
I’m more than happy to discuss ideas in this topic.
Research would become boring if we could do that 😜
April 25, 2025 at 4:35 PM
Research would become boring if we could do that 😜
Yes, I meant we don’t really know the ground truth…
Plus, the HRF differs across different regions of the brain and possibly across the population.
A very interesting but hard problem to solve!
With our algorithms we always saw that assuming the shape was good enough for most cases 🤷♂️
Plus, the HRF differs across different regions of the brain and possibly across the population.
A very interesting but hard problem to solve!
With our algorithms we always saw that assuming the shape was good enough for most cases 🤷♂️
April 25, 2025 at 4:34 PM
Yes, I meant we don’t really know the ground truth…
Plus, the HRF differs across different regions of the brain and possibly across the population.
A very interesting but hard problem to solve!
With our algorithms we always saw that assuming the shape was good enough for most cases 🤷♂️
Plus, the HRF differs across different regions of the brain and possibly across the population.
A very interesting but hard problem to solve!
With our algorithms we always saw that assuming the shape was good enough for most cases 🤷♂️
How would you learn the HRF from the data when the ground truth doesn’t exist?
Even if you were to transfer learn from task fMRI data, the model would be quite biased when applied to rest…
Even if you were to transfer learn from task fMRI data, the model would be quite biased when applied to rest…
April 25, 2025 at 2:32 PM
How would you learn the HRF from the data when the ground truth doesn’t exist?
Even if you were to transfer learn from task fMRI data, the model would be quite biased when applied to rest…
Even if you were to transfer learn from task fMRI data, the model would be quite biased when applied to rest…
The single echo is also available in AFNI as 3dPFM.
I would suggest using pySPFM with stability selection for better results though.
I would suggest using pySPFM with stability selection for better results though.
April 24, 2025 at 8:32 PM
The single echo is also available in AFNI as 3dPFM.
I would suggest using pySPFM with stability selection for better results though.
I would suggest using pySPFM with stability selection for better results though.
We showed that with our multi-echo multivariate paradigm free mapping approach we could get fairly close to GLM activity maps in task fMRI data. Stability selection (not multi-echo dependent) really helped us get robust results.
www.sciencedirect.com/science/arti...
www.sciencedirect.com/science/arti...
Whole-brain multivariate hemodynamic deconvolution for functional MRI with stability selection
Conventionally, analysis of functional MRI (fMRI) data relies on available information about the experimental paradigm to establish hypothesized model…
www.sciencedirect.com
April 22, 2025 at 9:39 PM
We showed that with our multi-echo multivariate paradigm free mapping approach we could get fairly close to GLM activity maps in task fMRI data. Stability selection (not multi-echo dependent) really helped us get robust results.
www.sciencedirect.com/science/arti...
www.sciencedirect.com/science/arti...
I did something similar to see my post stars a couple of months ago.
Such a fun application!
Such a fun application!
April 10, 2025 at 12:51 PM
I did something similar to see my post stars a couple of months ago.
Such a fun application!
Such a fun application!