In the framework of the “Fruits of Ethology” guest talk series we will have the pleasure to welcome our next speaker, Kauê M. Costa (Institute of Neurophysiology, Goethe University Frankfurt).
Date: 28th June, Thursday, 16.30
Location: ELTE, South Building, 7.110, Pázmány Péter sétány 1/c
Title: An alternative splicing event in dopamine neurons that selectively controls learning from disappointment
Learning from events that violate previously constructed expectations is a crucial ability for most animals. Midbrain dopamine neurons have been shown to represent errors between expected and experienced rewarding events: when an animal receives a better than expected reward, they burst, and when an animal receives a worse than expected rewards (i.e. is confronted with a negative prediction error), these neurons pause their activity. There is evidence that the duration of these pauses can be influenced by cell-autonomous properties, one of which is the A-type potassium current, which in certain dopamine neurons regulates the rebound of activity after inhibition.
In the present study, we aimed to elucidate if the modulation of dopamine neuron A-type currents by a particular regulatory channel subunit variant could, by its presumed effects on pausing, change learning from disappointment. In order to investigate this, we generated a transgenic mouse model in which we selectively removed from midbrain dopamine neurons an alternative splice variant of an A-type channel subunit named KChIP4a. We then trained these mice in an appetitive reinforcement learning task, which showed that selective KChIP4a removal from dopamine neurons strongly accelerated learning from reward omission (extinction learning), but did not impact learning from reward presentation (acquisition learning). Further analysis of the animals’ response dynamics revealed that this effect was due to a decrease in the probability of initiating goal directed actions, with no effect on action termination. Computational fitting of the behavioral data ratified that this phenotype could be attributed to a specific increase in the learning rate from negative prediction errors. In a subsequent behavioral phenotyping battery, we found no impact of KChIP4a removal on spontaneous behaviors and baseline cognitive properties, including locomotor activity, novel object preference, anxiety levels and working memory. In summary, we have identified a dopamine cell-type specific alternative splicing event that results in a specific acceleration of learning from disappointing events.
Márta Gácsi & Ádám Miklósi