Janik Goltermann
@janikgoltermann.bsky.social
CBT therapist and postdoc at Charité Berlin working in neuroimaging psychiatry🧠 Topics: childhood maltreatment, neurobiology, depression and replicability / open science
The journal is diamond open access, meaning free to read and free to publish
October 10, 2025 at 9:42 AM
The journal is diamond open access, meaning free to read and free to publish
🎓 Stay tuned for upcoming PhD position opening if you are interested to learn about depression, neuroimaging and causal inference
October 7, 2025 at 7:08 PM
🎓 Stay tuned for upcoming PhD position opening if you are interested to learn about depression, neuroimaging and causal inference
I'd be happy to but unfortunately, the availability of pediatric samples is limited at our lab. We are currently conducting a series of direct replications but mostly with adult samples so far. However, I would love to expand this to other labs and datasets if you are interested
September 18, 2025 at 5:53 AM
I'd be happy to but unfortunately, the availability of pediatric samples is limited at our lab. We are currently conducting a series of direct replications but mostly with adult samples so far. However, I would love to expand this to other labs and datasets if you are interested
Huge thanks to all involved, including: @smeinert.bsky.social @kflinkenfluegel.bsky.social @lisaleehr.bsky.social @jboehnlein.bsky.social @mariusgruber.bsky.social @kabro.bsky.social @vincenthammes.bsky.social @straubeb.bsky.social @nilsopel.bsky.social and everyone at @traplabms.bsky.social
September 16, 2025 at 4:21 PM
Huge thanks to all involved, including: @smeinert.bsky.social @kflinkenfluegel.bsky.social @lisaleehr.bsky.social @jboehnlein.bsky.social @mariusgruber.bsky.social @kabro.bsky.social @vincenthammes.bsky.social @straubeb.bsky.social @nilsopel.bsky.social and everyone at @traplabms.bsky.social
Our work adds to previous prominent contributions that question replicability in neuroimaging (e.g., @smarek0502.bsky.social and @tervoclemmensb.bsky.social in @nature.com - doi.org/10.1038/s415...)
Reproducible brain-wide association studies require thousands of individuals - Nature
Combined data from three large studies, with a total sample size of around 50,000 individuals, indicate that many previous studies linking the brain to complex phenotypes have been...
doi.org
September 16, 2025 at 4:21 PM
Our work adds to previous prominent contributions that question replicability in neuroimaging (e.g., @smarek0502.bsky.social and @tervoclemmensb.bsky.social in @nature.com - doi.org/10.1038/s415...)
The original finding of gray matter correlates of maltreatment has been published numerous time and is often described as established. Our consistent null results challenge that view and raise broader questions: how replicable are results in mental health neuroimaging?
September 16, 2025 at 4:21 PM
The original finding of gray matter correlates of maltreatment has been published numerous time and is often described as established. Our consistent null results challenge that view and raise broader questions: how replicable are results in mental health neuroimaging?
Similar results were seen when rerunning all analyses using parcellation-based measures of regional cortical thickness/surface and subcortical volumes
September 16, 2025 at 4:21 PM
Similar results were seen when rerunning all analyses using parcellation-based measures of regional cortical thickness/surface and subcortical volumes
Largest effects were found in models including HC and MDD groups while not controlling for diagnostic group (model 2 and 13) but disappeared when diagnosis was accounted for
September 16, 2025 at 4:21 PM
Largest effects were found in models including HC and MDD groups while not controlling for diagnostic group (model 2 and 13) but disappeared when diagnosis was accounted for
The significance of voxel overlap (i.e., replicability) was tested using permutations, resulting in no above-chance replicability across any pairwise or triple-wise cohort combination in any analysis (also when using p<.01 uncorrected)
September 16, 2025 at 4:21 PM
The significance of voxel overlap (i.e., replicability) was tested using permutations, resulting in no above-chance replicability across any pairwise or triple-wise cohort combination in any analysis (also when using p<.01 uncorrected)
Replicability was defined as the same voxel showing significance across cohorts. At p<.001 uncorrected (a liberal significance treshold for voxel-wise data) only two analyses resulted in minor overlap - k=2 and k=3 voxels. My personal favorite results item, Table 3:
September 16, 2025 at 4:21 PM
Replicability was defined as the same voxel showing significance across cohorts. At p<.001 uncorrected (a liberal significance treshold for voxel-wise data) only two analyses resulted in minor overlap - k=2 and k=3 voxels. My personal favorite results item, Table 3: