Phasic firing in dopaminergic neurons is sufficient for behavioral conditioning

Abstract

Natural rewards and drugs of abuse can alter dopamine signaling, and ventral tegmental area (VTA) dopaminergic neurons are known to fire action potentials tonically or phasically under different behavioral conditions. However, without technology to control specific neurons with appropriate temporal precision in freely behaving mammals, the causal role of these action potential patterns in driving behavioral changes has been unclear. We used optogenetic tools to selectively stimulate VTA dopaminergic neuron action potential firing in freely behaving mammals. We found that phasic activation of these neurons was sufficient to drive behavioral conditioning and elicited dopamine transients with magnitudes not achieved by longer, lower-frequency spiking. These results demonstrate that phasic dopaminergic activity is sufficient to mediate mammalian behavioral conditioning.

Fig. 1

Specific ChR2 expression in DA neurons. (A) Schematic of the Cre-dependent AAV; the gene of interest is doubly flanked by two sets of incompatible lox sites. Upon delivery into TH∷IRES-Cre transgenics, ChR2-EYFP is inverted to enable transcription from the EF-1α promoter. (B) Confocal images showing cell-specific ChR2-EYFP expression (green) in TH neurons (red). (C) Statistics of expression in TH neurons (n = 491); error bars represent SEM throughout. (D) Labeled VTA DA neurons project to downstream brain regions; confocal images of ChR2-EYFP–positive axons (green) innervating target neurons in NAc (NeuN, red).

Fig. 2

Photoactivation of DA neurons in intact tissue. (A) Recording from transduced neurons in acute VTA slices. Recorded neurons are verified by intracellular dye loading (ChR2-EYFP, green; AlexaFluor 594, red). (B) Continuous blue light (473 nm) evokes inward photocurrents. (C) Summary of photocurrent properties (n = 10). (D) Whole-cell recording of DA neurons showing spontaneous activity and tonic and phasic firing evoked by 1-Hz and 50-Hz light flash trains, respectively (25 flashes, 15 ms per flash). (E) Light-evoked spike trains are reliable over a range of frequencies; percentage of action potentials evoked by 25 light flashes at indicated frequencies (1 to 50 Hz) is shown (n = 7). (F) In vivo optrode recording of VTA DA neurons in a transduced TH∷IRES-Cre anesthetized mouse showing light-evoked DA spikes; see fig. S2 (Inset) Typical triphasic DA extracellular spike.

Fig. 3

Photoactivation of DA neurons induces place preference. (A) CPP timeline (see also fig. S3). (B) Light delivery parameters; 25 flashes at 1 or 50 Hz were delivered with a periodicity of 1 min. (C) Optical fiber is inserted through a cannula guide implanted over the VTA to photoactivate DA neurons. (D) Representative density maps showing conditioned preference; pseudocolor represents duration at each position. (E) Conditioning effect of DA neuron modulation. (Left) Comparison of time in each chamber during pretest (white) and posttest (gray). (Right) Comparison of preference scores for experimental (Phasic Stim, n = 13) and control cohorts (No Stim Control, n = 9; Littermate Control, n = 9). n.s. indicates not significant, *P < 0.05, **P < 0.01, and ***P < 0.001. (F) Difference scores (calculated as the difference between time spent during pre- and posttest in the specified chamber) for each chamber shows a statistically significant shift in place preference. (G) Analysis of anxiety (fractional time in center of a chamber; n = 13). (H) Chamber entries, long-range locomotion (different sequential beam breaks), and short-range locomotion (repeated break of the same beam) during pre- and posttest (n = 13).

Fig. 4

Phasic DA neuron stimulation leads to transient DA release and place preference. (A and B) Effect of phasic (50 Hz) (A) and tonic (1 Hz) (B) stimulation on place preference. (Left) Time spent in each chamber during preference test. (Right) Comparison of preference scores; for A, n = 7, and for B, n = 6. *P < 0.05. (C) Relative effects of stimulation frequency examined by using the difference scores for phasic, tonic, and nonassociative control (same as No Stim in Fig. 3E) cohorts. (D) FSCV measurements of VTA stimulation–triggered transient DA release in NAc in anesthetized TH∷IRES-Cre mice. (Top) Representative voltammetry traces during phasic (25 flashes per 50 Hz) and tonic (16 flashes per 1 Hz) stimulation of VTA. (Insets) Background-subtracted voltammogram taken from the peak of stimulation, indicating that signal measured is DA on the basis of comparison to voltammograms of DA obtained in vitro. (Bottom) All background-subtracted voltammograms recorded over the 20-s interval. y axis is applied potential (E_apps versus Ag/AgCl reference electrode); x axis is the time at which each voltammogram was recorded. Current changes at the electrode are encoded in color. DA can be seen during stimulation at the feature ∼0.650 V (oxidation peak encoded as green) and between ∼−0.20 V and ∼−0.25 V at the end of the voltage scan. (E) Comparison of phasic and tonic light-evoked DA transients (n = 3).