1
$\begingroup$

The human brain is a signal processing system. Input streams contain sensory, motor, reward and possibly more signals mixed. I wonder if neuro transmitters associated with reward increase or decrease memorization of those streams. That would result in a tendency to memorize streams co-occurring with high reward over those co-occurring with low reward. Thus, attractors would settle for better streams and we would perform actions optimizing the reward.

$\endgroup$
  • 1
    $\begingroup$ So are you asking for models that use neuro-transmitters associated with reward (such as dopamine) to drive learning? $\endgroup$ – Seanny123 Mar 16 '16 at 23:54
  • $\begingroup$ @Seanny123 That would be interesting as well. My question is whether there is biological evidence for faster learning or increased plasticity under e.g. dopamine emission. $\endgroup$ – danijar Mar 18 '16 at 20:21
2
$\begingroup$

The effect of dopamine signaling is one of the areas where computational neuroscience has provided insight into brain mechanisms, specifically via Reinforcement Learning (RL) models.

Based off this paper and PHD thesis from Dan Rasmussen, this first began with this publication from Schultz showing that dopamine acts as a reward signaler. From the abstract:

Dopamine neurons are activated by rewarding events that are better than predicted, remain uninfluenced by events that are as good as predicted, and are depressed by events that are worse than predicted. By signaling rewards according to a prediction error, dopamine responses have the formal characteristics of a teaching signal postulated by reinforcement learning theories.

To answer your question directly, it does seem the speed/plasticity of learning is modulated by dopamine neurons.

$\endgroup$
  • 2
    $\begingroup$ But also see Friston et al., 2012 for a more general theory of dopamine function $\endgroup$ – mrt Mar 19 '16 at 20:28
  • $\begingroup$ Thanks for your answer and sorry for the late reply. Do you conclude your last sentence from the cite above? I understand that dopamine is associate with reward. But that doesn't directly relate to whether high dopamine level means much learning and low dopamine level (or negative reward) means less learning, does it? $\endgroup$ – danijar May 11 '16 at 0:36
  • 1
    $\begingroup$ @danijar you're correct. I've accidentally jumped a level in my induction. I'll try and fix that. $\endgroup$ – Seanny123 May 11 '16 at 16:27
2
$\begingroup$

A bit of context

In mammalian brains, a main neurotransmiter linked with reward is Dopamine. This molecule is produced in the Ventral Tegmental Area (VTA) and in the Substantia Nigra (SN).

In a very famous study, Schultz recorded neurons in this two regions (at the time, we didn't really make the difference between them), and realised that the activity of these neurons was linked with reward. More precisely, neurons were coding for Reward Prediction Error, a feature that is implemented in Reinforcement Learning models.

Recently, some researchers decided to adress your question. To do this, they used optogenetics, wich is a technique that permit you to modulate the activity of neurons with light.

You can read the paper here : McNamara et al. Nat. Neurosci. (2014)

What are the main results ?

Mice with increased dopamine neuron's activity had better performance in a spatial task. They were better at learning a new maze. On top of that, the increased activity of VTA neurons induced a higher amount of specific oscillations (called Sharp Wave Ripples) in the Hippocampus, a structure important in spatial learning. We start to have strong evidence that this oscillations are important for memory, as disrupting them impair memory (see Jadhav et al. Science (2012) and Girardeau et al. Nat. Neuosci. (2009) if you are interested in the subject)

Note that this study was done in mice. Their is probably papers making a link between the activity of the VTA and learning using FMRi on humans, but I don't know the litterature enough.

$\endgroup$
  • $\begingroup$ Thanks a lot! Do you happen to know if there are any studies on whether domapine increases plasticity at the synapse level? I.e. is a synapse more likely to grow/strengthen under higher dopamine concentration? $\endgroup$ – danijar Jan 14 at 0:20
  • $\begingroup$ @danijar You're welcome ! There is indeed evidence that dopamine modify the strength of synapses. More precisely, dopamine modify Long Term Potentiation, one of the main mecanism known to modify the strength of synapses (One exemple I just found : S. Li et al Nat. Neurosci. (2003) . It is a really good question and I will edit my answer when I have time. $\endgroup$ – Hornycar Jan 14 at 11:22

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.