⁸This is perhaps oversimplifying the original idea: Recent research in neuroscience has emphasized that, as in the original definition above, it is important that cell A participates in the activation, i.e. is has a causal influence of cell B. This would usually mean that cell A is activated before cell B (Markram et al., 2012). However, in most implementations of Hebbian learning in AI, such causal and temporal aspects are not used. Actually, the details of how Hebbian learning works in the brain are not very well understood.