appearently headers work the other way around with gitlab markdown

README.md

-### Goal
+# Goal
 This project aims to extend the methods presented in 
 https://www.frontiersin.org/articles/10.3389/fncom.2017.00048/full
 
@@ -10,8 +10,8 @@ https://www.frontiersin.org/articles/10.3389/fncom.2017.00048/full
 The data/ subdirectory contains (for your convenience) connectomes from several other research projects. In no
 particular order these are:
 
-### Data sources
-# BBP
+# Data sources
+### BBP
 data/bbp/ contains data downloadable from bbp.epfl.ch. They wish to be cited by:
 1. Markram H, et al. (2015). Reconstruction and Simulation of Neocortical Microcircuitry. Cell 163:2, 456 - 492. doi: 10.1016/j.cell.2015.09.029
 
@@ -19,17 +19,17 @@ data/bbp/ contains data downloadable from bbp.epfl.ch. They wish to be cited by:
 
 3. Reimann MW, et al., (2015). An Algorithm to Predict the Connectome of Neural Microcircuits. Front. Comput Neurosci. 9:28. doi: 10.3389/fncom.2015.00120
 
-# C.Elegans
+### C.Elegans
 data/c.elegans/ contains (curated) data from the wormwiring.org project. Im not sure how they wish to be cited, but the
 data is from:
 https://wormwiring.org/pages/adjacency.html
 
-# q-rewiring
+### q-rewiring
 These are artificial connectomes of SNN trained via Q-rewiring (Horst Petschenig). See
 TBA
 for details.
 
-# deep-rewiring
+### deep-rewiring
 These are artifical connectoms of SNN trained via deep-rewiring to solve sequential MNIST. See
 G. Bellec, D. Kappel, W. Maass, and R. Legenstein. Deep rewiring: training very sparse deep networks. International Conference on Learning Representations (ICLR), 2018.