I’d like to thank everyone involved in the project! @romy-lorenz.bsky.social, @karolisdegutis.bsky.social, Daniel Haenelt, Robert Trampel, @nikweiskopf.bsky.social

We argue that our results are biologically plausible, considering both previous research and methodological aspects of layer fMRI acquisition and analysis. Consequently, the functional role of human dlPFC layers, especially deep layers, in working memory remains uncertain.

Using confidence interval estimation and a series of control analyses, we show that the failure to replicate was likely not an issue of low sensitivity, data quality or methodological choices, such as automatic segmentation and ROI selection.

We replicated the superficial layer involvement in working memory manipulation but found no evidence for deep layer-specific activation during motor response. Instead, we observed activation across both layers.

We used a fully automated, pre-registered analysis pipeline, that included automatic ROI definition by combining anatomical information (HCP MMP 1.0 atlas) with functional activation.

A seminal layer fMRI study (Finn et al. 2019) demonstrated that dlPFC layers were active during the manipulation of working memory items, whereas deeper layers were more engaged during motor responses.
www.nature.com/articles/s41...
@esfinn.bsky.social

I’d like to thank everyone involved in the project! @romy-lorenz.bsky.social, @karolisdegutis.bsky.social, Daniel Haenelt, Robert Trampel and @nikweiskopf.bsky.social

We argue that our results are biologically plausible, considering both previous research and methodological aspects of layer fMRI acquisition and analysis. Consequently, the functional role of human dlPFC layers, especially deep layers, in working memory remains uncertain.

Using confidence interval estimation and a series of control analyses, we show that the failure to replicate was likely not an issue of low sensitivity, data quality or methodological choices, such as automatic segmentation and ROI selection.

We replicated the superficial layer involvement in working memory manipulation but found no evidence for deep layer-specific activation during motor response. Instead, we observed activation across both layers.