Are PCA eigenvalues equivalent to semipartial correlation coefficients in linear regression?




I am a PhD student in my third month with a background in cognitive neuroscience. I loved statistics and thus developed a rather deep understanding of regression analyses.

Now, I started to apply computational models/analyses to the data I am currently looking at and the next step involves using PCA. My supervisor is talking a lot about how interesting it would be to see the component’s eigenvalues.

While trying to get my head around those, I found that they sound quite similar to what semipartial correlation coefficients in (linear) regression analyses express. Is that right? Can they even be viewed as being equivalent? Or am I misunderstanding something fundamental?

I hope this is the right place to ask questions like this, I was not able to find anything that appeared to be more suitable.

Thank you very much for any help!