Basolateral amygdala rapid glutamate release encodes an outcome-specific representation vital for reward-predictive cues to selectively invigorate reward-seeking actions

Abstract

Environmental stimuli have the ability to generate specific representations of the rewards they predict and in so doing alter the selection and performance of reward-seeking actions. The basolateral amygdala participates in this process, but precisely how is unknown. To rectify this, we monitored, in near-real time, basolateral amygdala glutamate concentration changes during a test of the ability of reward-predictive cues to influence reward-seeking actions (Pavlovian-instrumental transfer). Glutamate concentration was found to be transiently elevated around instrumental reward seeking. During the Pavlovian-instrumental transfer test these glutamate transients were time-locked to and correlated with only those actions invigorated by outcome-specific motivational information provided by the reward-predictive stimulus (i.e., actions earning the same specific outcome as predicted by the presented CS). In addition, basolateral amygdala AMPA, but not NMDA glutamate receptor inactivation abolished the selective excitatory influence of reward-predictive cues over reward seeking. These data support [corrected] the hypothesis that transient glutamate release in the BLA can encode the outcome-specific motivational information provided by reward-predictive stimuli.

Figure 1. Effect of basolateral amygdala AMPA or NMDA receptor inactivation on Pavlovian-instrumental transfer.

(A) Experimental procedure (see methods). CS, conditioned stimulus; O, outcome/reward; R, instrumental lever-press response; Ø, no reward delivery. During the PIT test both levers were available, but pressing was not reinforced and each CS was presented 4 times with intervening no-cue (Pre-CS) periods serving as a control. During CS presentation actions on the lever that, during training, earned the same outcome as the cue predicted were considered ‘same’ presses, while actions on the other available lever were considered ‘different’ (Diff) presses. (B) Schematic representation of microinfusion injector tips. Numbers to the lower right of each section represent the anterior-posterior distance (mm) from bregma of the section. Line drawings of coronal sections are taken from. (C,D) The Pavlovian-instrumental transfer effect; Lever-press rate (presses/min) averaged across levers during the control Pre-CS periods compared to pressing on lever that, in training, earned the same outcome as predicted by the CS (CS-Same) relative to pressing on the opposite lever (CS-Different) for the AMPA antagonist (C) or NMDA antagonist (D) group. Error bars +1 SEM. *p < 0.05, **p < 0.01, ***p < 0.001.

Figure 2. Experiment 2 Design.

(A) Schematic representation of the placement of the microelectrode array biosensor tips. Numbers to the lower right of each section represent the anterior-posterior distance (mm) from bregma of the section. (B) Testing procedures (see methods). CS, conditioned stimulus; O, outcome/reward; R, instrumental lever-press response; Ø, no reward delivery. During the PIT test both levers were available, but pressing was not reinforced and each CS was presented 4 times with intervening no-cue (Pre-CS) periods. During the CS presentation actions on the lever that, during training, earned the same outcome as the cue predicted were considered ‘same’ presses, while actions on the other lever were considered ‘different’ (Diff) presses. BLA glutamate concentration changes were continuously monitored with constant potential amperometry at glutamate biosensors during each test.

Figure 3. Basolateral amygdala glutamate release during instrumental conditioning.

(A) Representative glutamate concentration v. time trace during instrumental conditioning. Instrumental session started with lever insertion at time 0 s. Red asterisks represent significant transient glutamate concentration fluctuations above baseline (transients). Behavioral events are marked above the trace. (B) BLA glutamate transients that reached threshold were counted and averaged for each rat across the 2-min pre-test baseline period and the instrumental (inst) test sessions. (C) Glutamate transient amplitude (μM) was calculated as the peak amplitude of the transient minus the baseline concentration (first minima 0.5–5 s before the peak). (D) Lever press and food-port entry rate collapsed across the instrumental conditioning tests. (E) Between-subject correlation between glutamate transient frequency (Transients/min) and lever-press rate (Presses/min). Each rat is represented twice, once for each instrumental test. (F) Glutamate concentration v. time trace around initiating lever presses (occurring at time 0 s) averaged across all initiating presses in the session for a representative subject. Shading reflects +1 SEM across trials. (G) The likelihood of a glutamate transient in 10, 1-s bins, evenly distributed around presses divided by initiating presses (first press after the collection of an earned reward or after a >6 s pause in pressing, excluding presses that occurred within a pressing bout) or all lever presses combined (including both initiating and intra-bout presses), averaged across the two instrumental tests. The press occurs at time 0 s. Likelihood of a glutamate transient is defined as the percentage of presses (or initiating presses) that had a glutamate transient in the represented 1-s time bin. Asterisks represent significance relative to the control 1-s time bin, 5 s prior to the press. The raster plot shows the corresponding raw data with each subject represented on an individual line on the y-axis, tick marks represent the peak time of each significant glutamate transient surrounding initiating presses. Error bars +1 SEM. *p < 0.05, ***p < 0.001.

Figure 4. Basolateral amygdala glutamate release during Pavlovian-instrumental transfer.

(A) Average glutamate concentration change (μM) during the 2-min conditioned stimulus (CS) presentation (beginning at time 0 s) and immediately preceding 2-min pre-CS period averaged across trials during the PIT test for each rat and then averaged across rats. Dashed lines represent the between-subjects +1 SEM. (B) Glutamate transients that reached threshold were counted and averaged for each rat across the 2-min pre-test baseline period, the 2-min pre-CS periods and the 2-min CS periods. (C) Glutamate transient the amplitude (μM) was calculated as the peak amplitude of the transient minus the baseline glutamate concentration. (D) Lever-press rate (Presses/min) averaged across levers during the control Pre-CS periods compared to pressing during CS presentation distinguished by whether it was on lever that, in training, earned the same outcome as predicted by the presented CS (CS-Same Actions) relative to pressing on the opposite lever (CS-Diff Actions). Main effect of CS: F_2,14 = 8.85, p = 0.003. (E) Glutamate transient frequency (Transients/min) v. lever-press rate (Presses/min) between-subjects correlation. (F) Glutamate concentration v. time traces for the 5 s prior to and after initiating presses (occurring at time 0 s) during the CS averaged across all initiating presses for a representative subject. Shading reflects +1 SEM across trials. (G) The likelihood of a glutamate transient distributed in 1-s bins, 5 s prior to and after initiating presses (occurring at time 0 s; first press after a >6 s pause in pressing). Glutamate transient likelihood is defined as the percentage of initiating presses with a glutamate transient in the represented 1-s time bin. Asterisks represent significance relative to the first 1-s time bin. Raster plot displays corresponding raw data; each subject is represented on an individual line on the y-axis with tick color reflecting trial type. Tick marks represent the peak time of each glutamate transient that reached threshold surrounding initiating presses. Error bars +1 SEM. *p < 0.05, **p < 0.01, ***p < 0.001.

Figure 5. Outcome specificity of basolateral amygdala glutamate release during Pavlovian-instrumental transfer.

(A) Between-subjects correlation between glutamate transient frequency (Transients/min) and elevation [CS pressing/(CS + Pre-CS pressing)] v. response [CS-Same/(CS-Same + CS-Different)] ratio. (B) Glutamate concentration v. time traces for the 5 s prior to and after initiating presses (occurring at time 0 s) was averaged across all CS-Same initiating presses during the PIT test for the same representative subject as in 4F. Traces are divided by the type of outcome the action earned. Shading reflects the +1 SEM across trials. (C) The likelihood of a glutamate transient distributed in 1-s bins, 5 s prior to and after CS-Same initiating presses (occurring at time 0 s). Data are divided by the two CS-Same trial types. O1: Outcome 1 is defined as the outcome earned by the action that was exclusively (for 6/8 subjects) or preferentially (for 1/8 subjects) preceded by transient glutamate release. For the single subject in which he glutamate signal did not distinguish between outcome types outcome 1 was randomly assigned to the grain pellet outcome. Likelihood of a glutamate transient is defined as the percentage of initiating presses that had a glutamate transient in the represented 1-s time bin. Asterisks represent significance relative to the control 1-s time bin 5 s prior to the press. *p < 0.05, ***p < 0.001.