those graphs look about the same to me. if you told me these were generated from two seperate sets of draws from some normal distribution, I'd believe you.
if we're discussing loss of gray matter, shouldn't we be looking at just the differences pre-covid and post-covid? like how do we know the people with covid aren't just more likely to be brainlets?
>>13341273 >>13341294How can you pull apart the possibility that people that get COVID are just more likely to be brainlets? It says in the methodology:
> Before the COVID-19 pandemic, longitudinal (first- and second-timepoint scanning) had already begun in UKB imaging, with around 3,000 participants visiting for their second scan prior to scanning being paused in 2020 as a result of the pandemic. More recently, starting in February 2021, hundreds of UK Biobank participants who had already taken part in UKB imaging before the pandemic were invited back for a second scan. > Controls were then selected by identifying, from the remaining UK Biobank participants with negative resultsand note they paired each covid patient with a non-covid one by:
> ethnicity (white/non-white, as numbers were too low to allow for a finer distinction)and other criteria NOT including BMI and probably other important health characteristics.
They do say:
> For each IDP, we correlated the Scan 1 with Scan 2 values, separately for COVID patients and controls, resulting in two reproducibility measures (Pearson correlation r) for each IDP. This assumes that potential ageing and disease effects are subtle compared with between-subject variability and IDP noisewhich from my understanding means they did some linear regression between scan 1's and scan 2's and the line fits pretty well.
I don't get why they do this to the data. They have the pre and post-covid brain scans. Compare it to equivalent population same-ages brain scans.
