Prof. Gabriele Lohmann | Reproducibility, statistical inference and network analysis in fMRI at 3 Tesla and beyond

Guest Lecture

  • Date: Apr 10, 2017
  • Time: 05:00 PM - 06:00 PM (Local Time Germany)
  • Speaker: Prof. Gabriele Lohmann
  • Max Planck Institute for Biological Cybernetics and University Clinic, Tuebingen, Germany
  • Location: MPI for Human Cognitive and Brain Sciences
  • Room: Wilhelm Wundt Room (A400)
In recent years there has been a growing concern about the reproducibility of results obtained in human brain mapping using fMRI. Two major factors contribute to this problem, namely inflated false positive rates (Eklund et al, 2016) and lack of statistical power, i.e. inflated false negative rates (Button et al, 2013). While the detrimental effect of false positives is widely recognized, it may be less well known that inflated false negative rates may also diminish reproducibility. In this talk, I will propose a new algorithm called "LISA" that addresses this issue. Furthermore, LISA is applicable to ultrahigh-field fMRI data (>= 7T). It thus fills a gap as there are currently no statistical inference methods available that specifically target such data.

In the second part of the talk I will address network analysis in fMRI. I will introduce a new algorithm that identifies edges in a brain network that differentially respond to task onsets. The algorithm is called "Task-related Edge Density" (TED) (Lohmann et al, 2016). I will present results of TED applied to task-based fMRI data acquired at 9.4 Tesla.

Poster
Go to Editor View