Cognitive Learning Agent

Source code of the cognitive learning agent for emergent attention from basal ganglia onto the visual system from Maith, Schwarz & Hamker (2021).

Authors:

Visual system model is based on Beuth (2019)
Basal Ganglia model is based on Villagrasa et al. (2018).

Result	analysis	simulation	comment
input stimuli	-	`python create_stim.py X`	generates the random stimuli which are used in all simulations X = σ_V1 1 - σ_V1 = 22° 2 - σ_V1 = 30° 3 - σ_V1 = 14°
trial_BG.svg (Fig. 3)	`python trial_BG.py 1 0`	`python run_cla_one_trial.py`	simulation actually simulates two trials, second trial only for negative SNc response
Training.svg (Fig. 6), Training.txt (Training statistics)	`python Training.py`	`python run_cla_Training X 0`	X = simulaion IDs (e.g. 60 different simulations)
T2.svg (Fig. 7), T2.txt (T2 statistics)	`python T2.py`	`python run_cla.py X Y`	X = simulation IDs, Y = stimulus ID (0-T1, 1-T1-reversed)
T1_performance.txt (T1 performance)	`python T1_performance.py`	`python run_cla.py X Y`	X = simulation IDs, Y = stimulus ID (0-T1, 1-T1-reversed) like T2.py
Learn_BG.svg (Fig. 4)	`python Learn_BG.py`	`python run_cla.py X 0`	X = simulation IDs, for analyses use simulations which learned 65° gain
Learn_PFC.svg (Fig. 5)	`python Learn_PFC.py`	`python run_cla.py X 0`	X = simulation IDs, for analyses use simulations which learned 65° gain
EffectPFC.svg (Fig. 8)	`python EffectPFC.py`	`run_parallel_EffectPFC.sh`	in folder EffectPFC

python rename.py: change simulation IDs of specified files and copy files in new folder
run_parallel.sh: run python scripts parallel, multiple times