The dopaminergic neurons of the ventral tegmental area and substantia nigra pars compacta, located within the ventral mesencephalon, encode perhaps one of the most important signals for reinforcement learning in the brain: reward prediction error. This signal is encoded by the firing pattern of dopaminergic neurons, which controls the release of dopamine at target regions. Specifically, transient, impulse-dependent release of dopamine, driven by bursts of action potentials, is critical for natural processing in the brain. Just as critical, pauses in dopaminergic cell activity have opposite psychological meaning for reward information coding and are thought to signal the absence of an expected reward. It is likely the essential nature of this signal that connects disruptions of dopamine function to many of the symptoms of a wide range of psychiatric diseases, drug addiction, and in the extreme case of the degeneration of these cells, to Parkinson's Disease, including many of its cognitive aspects. Identification of the input pathways responsible for bursts and pauses is a key step in understanding the mechanism of reinforcement learning, but has so far proven elusive. This is largely due to the difficulty in accurately duplicating bursts and pauses under controlled experimental conditions such as those attainable during in vitro experiments. A second difficulty has been the inability to selectively activate identified afferents to dopaminergic neurons during controlled in vitro experiments.
The specific aims i n this proposal are designed to investigate the synaptic mechanisms by which specific identified afferents induce bursts and pauses, and how psychostimulants alter the input from those same afferents. This will identify the circuit basis and synaptic origin of the reward prediction error signal, and provide a mechanistic understanding of how drugs of abuse alter the reward prediction error signal. To achieve this, we will use recordings in conductance-clamp, where specific conductances can be imposed in a defined manner directly on dopaminergic neurons, along with selective in vitro stimulation of specific afferents following prior viral infection in vivo with channelrhodopsin (ChR2). The ChR2 strategy provides a clear advantage over simultaneous activation of all afferents by electrical stimulation since it will selectively identify the afferents responsible for bursts and pauses, in addition to those afferents with affected synaptic input after in vivo cocaine exposure.
The dopaminergic neurons located within the ventral mesencephalon encode perhaps one of the most important signals for reinforcement learning in the brain: reward prediction error. This signal is encoded by the firing pattern of dopaminergic neurons, which controls the release of dopamine at target regions, and it is likely the essential nature of this signal that connects disruptions of dopamine function to many of the symptoms of a wide range of psychiatric diseases, drug addiction, and in the extreme case of the degeneration of these cells, to Parkinson's Disease. This proposal aims to investigate the synaptic mechanisms by which specific afferents trigger the reward prediction error signal, and how drugs of abuse alter the input from those same afferents.
