Inhibitory neurons exhibit high controlling ability in the cortical microconnectome
Fig 3
Four processing steps as a combination of cell categorization and connectivity estimation: ① First, we calculated E-I bias (SLTE at the delay where TE shows the peak value). ② Then, we determined the sign, positive or negative by observing the clustered distributions in the two-dimensional space of the the 1st component (∝ ∑connections in top90% about peakTE SLTE) and the 2nd SLTEs components (∝ log10 FR). ③ We calculated Strength and Sharpness from the TE(d), Transfer Entropy as a function of delay d, according to the analysis scheme shown in Fig 2C and 2D. ④ Finally, we separated the Strength and Sharpness matrices into excitatory and inhibitory groups based on the classification of step ②. (b) The cell categorization preceded the connectivity determination because we should prepare the shuffled null data separately for excitatory and inhibitory synaptic connections to determine connected neuron pairs [Refer to the supplemental material of Shimono, Beggs, 2014 for more detailed information]. We separately produced the connectivity matrix for excitatory and inhibitory neurons by selecting strong or sharp connections as compared with the null shuffled data.