Section: New Results

The limbic loop

We explored the limbic loop by describing a series of neural mechanisms that propose how responding conditioning results from interactions between the amygdala, the nucleus accumbens and the limbic pole of the frontal cortex. In our models [1], this learning is also fed by exchanges with the hippocampus (episodic memory) and the sensory cortex (semantic memory) and we studied the major role of acetylcholine in these exchanges. This also allowed us to address the difficult question of the articulation between the respondant and operant conditioning in particular in the nucleus accumbens. We proposed an original mechanism whereby noradrenaline could modulate the balance between exploration and exploitation [12] based on an assessment of the level of uncertainty and its impact on performance [13].

Also in connection with this loop, we studied the dynamics of dopamine release in the midbrain, considered to play an essential role in the coding of the prediction error. This model developed in the framework of our collaboration with India (cf. § 9.3) proposes to introduce into the classical circuit, new actors (such as the pedunculopontine tegmental nucleus in the brainstem) and new functions (dissociation of amplitude and timing of the reward), that we will seek to corroborate in the future.

Lastly, we carried out a thorough study about the behavior of our model of associative memory in the hippocampus [23], and particularly about its resistance to interference.