Lateral Habenula Manipulation Influences the Use of Outcomes to Guide Probabilistic Choice

The burgeoning field of neuroeconomics has contributed valuable information to behavioral economics and microeconomic theory by elucidating the neural substrates utilized by humans and other organisms to make various types of cost/benefit decisions.  Neuroscience of decision-making has been heavily influenced by the discovery that individual dopamine neurons in the ventral tegmental area (VTA) encode reward prediction errors which aid in learning and making decisions in the face of uncertain outcomes.  Recently, the lateral habenula (LHb), has emerged as a key input to dopamine neurons in the (VTA), providing indirect inhibition to these neurons.  In opposition to the traditional reward prediction error signal represented by the VTA, LHb neurons reduce activity in response to rewards, and show phasic increases in firing after reward omission.  While its role in processing reward has become better understood, little is known about its relevance to choice behavior in cost/benefit decision scenarios.  The current set of experiments aimed to examine the role of the LHb specifically in probabilistic decision-making.  Rats were trained on a probabilistic discounting task in which animals chose between a small, certain reward and a large, uncertain reward.  Across a session, the probability of the risky option changed, becoming the less advantageous option long-term.  We aimed to understand the role of the LHb on this form of decision making through general tonic activity as well as task-related phasic activity.  To understand the contribution of general tonic LHb activity, we implanted rats with guide cannulae into the LHb.  Rats received counterbalanced infusions of saline or GABA agonists baclofen and muscimol to temporarily inactivate the region during a test session.  To assess the role of phasic task-relevant bursts of LHb activity, we implanted a stimulating electrode into the LHb.  On different test days rats received a pulse train in conjunction with various task components: after a risky win, after any risky choice, after a safe reward, before choice, and during the inter-trial interval.  Our results suggest that manipulating phasic activity of the LHb influences probabilistic choice.  Particularly, stimulation of the LHb after a risky win decreases choice of this option.  This manipulation decreases win-stay tendencies, shifting subsequent choice away from a risky choice with a successful outcome on a previous trial.  We hypothesize that this is due to a blunting of reward sensitivity; when delivered at the right moment, this can bias choice away from a risky option, that while recently lucky, may be disadvantageous in the long-term.