Optogenetic Activation of the Basolateral Amygdala Promotes

Research Articles: Behavioral/Cognitive Optogenetic activation of the basolateral amygdala promotes both appetitive conditioning and the instrumental pursuit of reward cues https://doi.org/10.1523/JNEUROSCI.2196-19.2020

Cite as: J. Neurosci 2020; 10.1523/JNEUROSCI.2196-19.2020 Received: 11 September 2019 Revised: 6 January 2020 Accepted: 8 January 2020

This Early Release article has been peer-reviewed and accepted, but has not been through the composition and copyediting processes. The final version may differ slightly in style or formatting and will contain links to any extended data.

Alerts: Sign up at www.jneurosci.org/alerts to receive customized email alerts when the fully formatted version of this article is published.

1 Optogenetic activation of the basolateral amygdala promotes both appetitive conditioning and 2 the instrumental pursuit of reward cues 3 4 Running title: Basolateral amygdala and reward cue properties 5 6 7 Alice Servonneta, Giovanni Hernandeza,b, Cynthia El Hagec, Pierre-Paul Rompréa and Anne-Noël 8 Samahac,d 9 10 aDepartment of Neurosciences, Faculty of Medicine, Université de Montréal; 2900 Edouard-Montpetit 11 boulevard, Montreal (H3T 1J4), Quebec, Canada; bDepartment of Psychiatry, Douglas Mental Health 12 University Institute, McGill University; 6875 Boulevard Lasalle, Montreal (H4H 1R3), Quebec, Canada; 13 cDepartment of Pharmacology and Physiology, Faculty of Medicine, Université de Montréal; 2900 14 Edouard-Montpetit boulevard, Montreal (H3T 1J4), Quebec, Canada; dGroupe de recherche sur le 15 système nerveux central, Faculty of Medicine, Université de Montréal; 2900 Edouard-Montpetit 16 boulevard, Montreal (H3T 1J4), Quebec, Canada 17 18 19 CORRESPONDING AUTHOR 20 21 Anne-Noël Samaha, PhD 22 Department of Pharmacology and Physiology 23 Université de Montréal 24 C.P. 6128, Succursale Centre-ville 25 Montreal, Quebec 26 H3C 3J7 27 Canada 28 Tel: 514 343 6111 x. 32788 29 Fax: 514 343 2291 30 [email protected] 31 32 Number of pages: 35 33 Number of figures: 7 34 Number of words for Abstract: 241 35 Number of words for Introduction: 626 36 Number of words for Discussion: 1137 37 38 CONFLICT OF INTEREST 39 40 ANS was a scientific consultant for H. Lundbeck A/S as this research was being carried out. This has 41 had no influence on the work. The remaining authors report no biomedical financial interests or 42 potential conflicts of interest. 43 44 ACKNOWLEDGEMENTS 45 46 This work was supported by grants from the National Science and Engineering Research Council of 47 Canada (grant number 355923) and the Canada Foundation for Innovation to ANS (grant number 48 24326). ANS holds a salary award from the Fonds de la Recherche du Québec-Santé (grant number 49 28988). We thank Nadia Chaudri, Franca Lacroix, Franz Villaruel, Ivan Trujillo-Pisanty, Jonathan Britt, 50 Sean J. Reed, Mike J.F. Robinson and Kent C. Berridge for generous advice on implementation of in 51 vivo optogenetics procedures in our laboratory. We also thank Ellie-Anna Minogianis for technical 52 support. 53

54 ABSTRACT

56 Reward-associated stimuli can both evoke conditioned responses and acquire reinforcing

57 properties in their own right, becoming avidly pursued. Such conditioned stimuli (CS) can guide

58 reward-seeking behaviour in adaptive (e.g., locating food) and maladaptive (e.g., binge eating) ways.

59 The basolateral amygdala (BLA) regulates conditioned responses evoked by appetitive CS, but less is

60 known about how the BLA contributes to the instrumental pursuit of CS. Here we studied the influence

61 of BLA neuron activity on both behavioural effects. Water-restricted male rats learned to associate a

62 light-tone cue (CS) with water delivery into a port. During these Pavlovian conditioning sessions, we

63 paired CS presentations with photo-stimulation of channelrhodopsin-2 (ChR2)-expressing BLA

64 neurons. BLA photo-stimulation potentiated CS-evoked port entries during conditioning, indicating

65 enhanced conditioned approach and appetitive conditioning. Next, new rats received Pavlovian

66 conditioning without photo-stimulation. These rats then received Instrumental conditioning sessions

67 where they could press an inactive lever or an active lever that produced CS presentation, without

68 water delivery. Rats pressed more on the active versus inactive lever, and pairing CS presentation

69 with BLA-ChR2 photo-stimulation intensified responding for the CS. This suggests that BLA-ChR2

70 photo-stimulation enhanced CS incentive value. In a separate experiment, rats did not reliably self-

71 administer BLA-ChR2 stimulations, suggesting that BLA neurons do not carry a primary reward signal.

72 Lastly, intra-BLA infusions of d-amphetamine also intensified lever-pressing for the CS. The findings

73 suggest that BLA-mediated activity facilitates CS control over behaviour by enhancing both appetitive

74 Pavlovian conditioning and instrumental pursuit of CS.

76 SIGNIFICANCE STATEMENT

78 Cues paired with rewards can guide animals to valuable resources such as food. Cues can

79 also promote dysfunctional reward-seeking behaviour, as in over-eating. Reward-paired cues

80 influence reward seeking through two major mechanisms. First, reward-paired cues evoke conditioned

81 anticipatory behaviours to prepare for impending rewards. Second, reward-paired cues are powerful

82 motivators and they can evoke pursuit in their own right. Here we show that increasing neural activity

83 in the basolateral amygdala enhances both conditioned anticipatory behaviours and pursuit of reward-

84 paired cues. The basolateral amygdala therefore facilitates cue-induced control over behaviour by

85 both increasing anticipation of impending rewards and making reward cues more attractive.

87 INTRODUCTION

89 Initially neutral cues (sights, sounds or places) that predict rewards such as food and water

90 exert powerful control over behaviour. For instance, reward-paired cues [or conditioned stimuli (CS)]

91 can acquire incentive motivational value (Bolles, 1972; Bindra, 1978), thereby ‘goading an individual

92 into action’ (Flagel et al., 2009). In this regard, CS can i) elicit approach and attention, allowing

93 animals to prepare for impending rewards (Hearst and Jenkins, 1974), ii) energize ongoing reward-

94 seeking behaviours (Rescorla and Solomon, 1967), iii) trigger reinstatement of extinguished reward-

95 seeking behaviour (de Wit and Stewart, 1981), and iv) reinforce learning of new instrumental

96 behaviours (Mackintosh, 1974; Cardinal et al., 2002). Through these effects, CS guide behaviour

97 towards rewards necessary for survival. However, changes in the response to CS can contribute to

98 excessive reward-seeking behaviours (as in addiction) or conversely, low levels of appetitive

99 behaviour (as in depression).

100

101 Prior studies have examined the role of the basolateral amygdala (BLA) in the capacity of CS

102 to both evoke conditioned approach and influence instrumental behaviour. BLA lesions (Burns et al.,

103 1993) and optogenetic stimulation of BLAÆnucleus accumbens shell neurons (Millan et al., 2017)

104 both attenuate CS-evoked conditioned responses. Similar effects are seen with optogenetic inhibition

105 of either BLA neurons expressing the Ppp1r1b gene (Kim et al., 2016) or BLAÆnucleus accumbens

106 core neurons (Stuber et al., 2011). The BLA is also thought to be necessary for the expression of CS-

107 controlled instrumental behaviour. Decreasing BLA function with lesions (Cador et al., 1989; Everitt et

108 al., 1991; Brown and Fibiger, 1993; Burns et al., 1993; White and McDonald, 1993; McDonald and

109 Hong, 2004; McDonald et al., 2010), pharmacological agents (Grimm and See, 2000; Kantak et al.,

110 2002; McLaughlin and See, 2003; Rogers et al., 2008; Gabriele and See, 2010) or optogenetic

111 methods (Stefanik and Kalivas, 2013) suppresses CS-controlled instrumental behaviour. However,

112 these studies used tasks that potentially confound the motivational effects of the CS and the

113 unconditioned stimulus (UCS), and/or neuronal manipulation methods that do not allow control of

114 neural activity coincident with CS occurrence (e.g., lesions/pharmacological agents).

115

116 In this context, key questions remain. First, how does increased BLA-mediated neuronal

117 activity during CS presentation influence respectively, CS-evoked conditioned behaviours and the

118 instrumental pursuit of CS? BLA neurons fire in response to CS presentations during appetitive

119 conditioning (Tye and Janak, 2007; Ambroggi et al., 2008; Tye et al., 2008). The functional

120 significance of this is not fully understood. Second, CS can motivate behaviour through many

121 dissociable psychological processes (Cardinal et al., 2002), what processes might BLA-dependent

122 activity regulate? Increased BLA activity might mediate the specific incentive value attributed to the

123 CS. If so, then increased BLA activity should alter CS motivational properties preferentially when it is

124 explicitly paired with CS presentations. The BLA might also arouse a general motivational state,

125 thereby ‘setting the occasion’ to perform a CS-controlled goal-directed behaviour (Lajoie and Bindra,

126 1976; Rescorla, 1988). If so, then increased BLA activity should alter CS incentive value, even when

127 increased BLA activity is explicitly unpaired with CS presentations.

128

129 We addressed these questions using in vivo optogenetics combined with Pavlovian and

130 Instrumental conditioning procedures. First, we determined whether photo-stimulation of BLA neurons

131 is intrinsically rewarding, as assessed by self-stimulation behaviour. We compared self-stimulation of

132 BLA neurons with self-stimulation of adjacent central amygdala (CeA) neurons, as rodents will self-

133 stimulate into the CeA (Seo et al., 2016; Baumgartner et al., 2017; Kim et al., 2017). Second, we

134 determined how photo-stimulation of BLA neurons influences appetitive conditioned responses, as

135 assessed by CS-evoked approach behaviour that indicates expectation of the primary reward

136 (Tolman, 1932; Hearst and Jenkins, 1974). Finally, we assessed how photo-stimulation of BLA

137 neurons influences CS-controlled instrumental behaviour, by measuring the capacity of a CS to

138 support the spontaneous learning of a new instrumental behaviour (Mackintosh, 1974; Robbins, 1978;

139 Cardinal et al., 2002).

140

141 METHODS

142

143 Animals

144

145 Male Sprague-Dawley rats (Charles River Laboratories, Montreal, Canada; 200-275 g on

146 arrival) were housed individually on a 12-h light/dark cycle (lights off at 8:30 a.m.). They were tested

147 during the dark phase of the circadian cycle. Food and water were available ad libitum, except in

148 Experiments 3-4, where water access was restricted to 2 h/day. This was to facilitate Pavlovian

149 conditioning using water as the UCS (see below for details). The Université de Montréal approved all

150 procedures involving animals and procedures followed the guidelines of the Canadian Council on

151 Animal Care.

152

153 Intra-cerebral surgery

154

155 Rats weighing 325-375 g were anesthetized with isoflurane and placed on a stereotaxic

156 apparatus. For photo-stimulation of amygdala neurons in Exps. 1-3, rats received bilateral infusions of

157 AAV5-hSyn1-hChR2(H134R)-eYFP (provided by Dr. Karl Deisseroth; UNC Vector Core, NC, USA)

158 into either the BLA (mm relative to Bregma: AP -2.8, ML ± 5.0; mm relative to skull surface: DV -8.4)

159 or CeA (mm relative to Bregma: AP -2.6, ML ± 4.3; mm relative to skull surface: DV -7.9). Control rats

160 received an optically inactive AAV-eYFP virus (AAV5-hSyn1-eYFP, UNC Vector Core). The hSyn

161 promoter is neuron-specific and allows gene expression in both excitatory and inhibitory neurons

162 (Dittgen et al., 2004). Using a glass pipette (tip diameter of ~50 μm) coupled to a Nanoject II

163 (Drummond Scientific, PA, USA), we administered 27 microinjections of 36.8 nL each (23 nL/second,

164 at 10-second intervals; total volume of ∼1 μL/hemisphere) into each brain region. After the infusions,

165 the glass pipette was left in place for 10 more minutes. In Exp. 4, guide cannulae (26 GA, model

166 C315G, HRS Scientific, Montreal, Canada) were implanted 2 mm dorsal to the BLA (mm relative to

167 Bregma: AP -2.4, ML ± 5.5; mm relative to skull surface: DV -6.6) or dorsal to the amygdala, without

168 targeting the BLA specifically, as a neuroanatomical control (referred to as ‘Amygdala’; mm relative to

169 Bregma: AP -2.3, ML ± 5.1; mm relative to skull surface: DV -6.2). In Exp. 1, the craniectomy was

170 sealed with bonewax (Ethicon, NJ, USA). In Exps. 2-3, an optic fiber implant (∼300 μm core diameter,

171 numerical aperture of 0.39; ThorLabs, NJ, USA; glued with epoxy to a ferrule, model F10061F340,

172 Fiber Instrument Sales Inc., Oriskany, NY, USA) was implanted in each hemisphere, 0.2 mm dorsal to

173 the virus injection site. Four to 6 stainless steel screws were then anchored to the skull, and optic fiber

174 implants or cannulae were fixed with dental cement. Optic fiber implants were protected with a sleeve

175 and a dummy. Guide cannulae were sealed with obturators (model C315CD, HRS Scientific).

176 Optogenetic manipulations started at least 4 weeks following virus injection, to allow sufficient viral

177 expression (Zhang et al., 2010).

178

179 In vivo electrophysiology

180

181 We used in vivo electrophysiology to confirm laser-induced action potentials in ChR2-

182 expressing neurons in the BLA and CeA. Anesthetized rats (urethane, 1.2 g/kg, i.p.) were placed

183 inside a Faraday cage on a stereotaxic frame equipped with a body temperature controller. Optrodes

184 were implanted above the BLA and CeA. Optrodes were constructed using an extracellular Parylene

185 coated tungsten electrode (1 MΩ, ∼125 μm outer diameter; FHC, Bowdoin, ME, USA) glued with

186 epoxy to an optical fiber (∼300 μm core diameter, numerical aperture of 0.39) with a ∼0.5 mm offset to

187 ensure illumination of recorded neurons. A reference electrode (insulated silver wire, 0.25 mm

188 diameter) was lowered into the back of the brain close to the cerebellum. The optrode and reference

189 electrodes were fixed with stainless steel screws anchored to the skull and bee wax. The optrode was

190 lowered by hydraulic microdrive into the BLA or the CeA to record single action potentials elicited by

191 laser stimulation (465 nm blue diode laser). Optrodes were linked to the laser via patch-cords built as

192 described in Trujillo-Pisanty et al. (2015).

193

194 The signal recorded from each optrode was fed into a high impedance headstage connected to

195 a microelectrode amplifier (Model 1800, A-M Systems, WA, USA). During photo-stimulation, the low-

196 and high-pass filters were set at 300 Hz and 5 kHz, respectively. To reduce the possibility of

197 photoelectric artifacts, we grounded the laser head and patch-cord. Action potentials were displayed

198 on an oscilloscope (Tektronix, Model TDS 1002, OR , USA). The signal was digitalized and stored

199 using DataWave recording (USB 16 channels) and DataWave SciWorks Experimenter Package

200 (DataWave Technologies, CO, USA).

201

202 Pavlovian Conditioning

203

204 Training and testing took place in standard operant chambers (Med Associates, VT, USA)

205 where a fan and a house-light were on. Rats had restricted water access for at least 3 days (2 h/day).

206 Starting on the next day, they were trained to associate a light-tone cue (see Fig. 1A; CS) with water

207 delivery (UCS; 100 μL) into a recessed receptacle, using Pavlovian conditioning procedures. The light-

208 tone cue consisted of illumination of two discrete lights for 5 seconds, combined with the extinction of

209 the house-light. This was immediately followed by an 1800-Hz, 85-dB tone. The tone lasted 0.18

210 seconds and was coincident with water delivery. The CS-UCS were presented on a variable interval of

211 60 seconds, 20 or 30 times/session. To determine the extent to which rats learned the CS-UCS

212 contingency, we measured CS-evoked conditioned approach behaviour. To this end, we quantified the

213 number of nose-pokes into the recessed water receptacle during each 5-second light cue presentation

214 (conditioned stimulus response; CSR) versus during the 5-second period preceding each CS

215 presentation (pre-conditioned stimulus response; PCSR). We computed a CSR/PCSR ratio for each

216 animal, on each conditioning session.

217

218 Instrumental Conditioning

219

220 To assess the capacity of the CS to control instrumental behaviour, we determined whether

221 after Pavlovian CS-UCS conditioning, rats would spontaneously learn a new instrumental response

222 (lever-pressing) to earn CS presentations, without the UCS. This procedure dissociates incentive

223 motivation for the CS versus for the UCS, because the instrumental response is new and not

224 previously reinforced by the UCS (Mackintosh, 1974; Robbins, 1978; Cardinal et al., 2002). First, rats

225 were placed in the operant chambers for a lever habituation session, during which they could sample

226 two test levers for the first time. As shown in Fig. 1B, pressing the active lever produced the CS,

227 without water delivery, according to a random-ratio 2 (RR2) schedule. Pressing on the active lever

228 during CS presentation or on the inactive lever had no programmed consequences but was recorded.

229 The lever habituation session ended after 10 active lever presses or 40 minutes. To measure the

230 incentive motivational value of the CS, rats received additional instrumental test sessions. During

231 these sessions, conditions were the same as during the lever habituation session, except that lever

232 presses were not limited. Sessions ended after 20 or 40 minutes. We refer to these sessions as

233 ‘operant responding for the CS’.

234

235 Exp. 1: Effects of photo-stimulation on action potentials in ChR2-expressing BLA and CeA

236 neurons in vivo

237

238 As shown in Fig. 2A, rats received either the ChR2-eYFP (n = 4) or eYFP (n = 1) virus into the

239 BLA of one hemisphere and into the CeA of the contralateral hemisphere. At least 4 weeks later, rats

240 were anesthetized and in vivo neuronal firing was measured following photo-stimulation [squared light

241 pulses of 5 ms delivered at 1, 10, 20 or 40 Hz at 10 mW; based on Huff et al. (2013) and Robinson et

242 al. (2014)]. These are the photo-stimulation parameters used in the behavioural studies below, with

243 frequencies ≤ 20 Hz, at which we observed excellent ChR2 fidelity. Importantly, BLA neurons also fire

244 in vivo at frequencies ≤ 20 Hz in behavioural tasks involving reward cues (Tye and Janak, 2007;

245 Ambroggi et al., 2008; Tye et al., 2008).

246

247 Exp. 2: Effects of photo-stimulating ChR2-expressing BLA or CeA neurons on lever-pressing

248 behaviour

249

250 If photo-stimulation of BLA neurons is intrinsically rewarding, it could reinforce lever pressing

251 behaviour and this would confound interpretation of subsequent results. Thus, here we determined

252 whether otherwise naïve rats would reliably lever press for photo-stimulation of BLA. We also

253 evaluated self-stimulation of CeA neurons, as photo-stimulation of CeA ChR2 has been reported to

254 sustain self-stimulation (Seo et al., 2016; Baumgartner et al., 2017; Kim et al., 2017). As shown in Fig.

255 3A, rats received bilateral injections of the ChR2-eYFP or eYFP virus into the BLA or CeA.

256 Experimental rats were ChR2-eYFP rats (n = 5/subregion) allowed to lever press for photo-stimulation.

257 Control rats included i) rats expressing ChR2-eYFP in the BLA (n = 3) or CeA (n = 2) that could lever

258 press but this did not produce photo-stimulation, and ii) rats expressing eYFP in the BLA (n = 3) or

259 CeA (n = 2) and allowed to lever press for photo-stimulations. Throughout the study, lever-pressing

260 behaviour was similar across control groups. Thus, they were pooled together for final analysis (n =

261 10). Photo-stimulation was bilateral except where noted otherwise.

262

263 As shown in Fig 3A, the rats were allowed to press a lever to obtain a 5.18-second laser

264 stimulation (20 Hz frequency, unless stated otherwise) paired with a 5.18-second presentation of the

265 light-tone stimulus described above. Importantly, rats were previously naïve to the light-tone stimulus,

266 such that this stimulus had not previously been paired with water or any other outcome in these rats.

267 During all sessions, active lever presses during photo-stimulation and inactive lever presses had no

268 programmed consequences, but both were recorded. Daily sessions ended after self-administration of

269 30 stimulations or 30 minutes, unless stated otherwise. First, for at least 2 sessions (1 session/day),

270 pressing the active lever produced photo-stimulation under a fixed-ratio of 1 (FR1) schedule of

271 reinforcement. The rats were then tested under RR2 and RR4 schedules, with 2 sessions/schedule.

272 Then, rats were given 2 sessions where photo-stimulation was available under a progressive ratio 5

273 schedule of reinforcement (PR5). During these sessions, the number of active lever presses required

274 to earn each successive photo-stimulation increased by a factor of 5, and sessions ended after 30

275 stimulations or 30 minutes (Rossi et al., 2013). Extinction responding was then evaluated during two

276 40-minutes sessions, based on Ilango et al. (2014). During minutes 0 to 5 and 20 to 25 of the

277 extinction sessions, lever pressing was reinforced with photo-stimulation under RR2. For the

278 remaining minutes of each session, lever pressing produced the light-tone stimulus, without photo-

279 stimulation. At the 20-minute mark, a single, non-contingent photo-stimulation combined with the tone-

280 light cue indicated that photo-stimulation was available once again. Next, we assessed the influence of

281 laser stimulation frequency on lever pressing behaviour during 3 sessions (5, 10 and 20 Hz, one

282 frequency/session/day, counterbalanced). We then assessed reversal learning for 2 sessions during

283 which the active and inactive levers were switched. If photo-stimulation of BLA or CeA neurons is

284 reinforcing, then ChR2-BLA rats and ChR2-CeA rats should stop responding on the newly non-

285 reinforced lever, and increase responding on the newly reinforced lever. Lastly, the rats were given a

286 final test session to determine whether unilateral photo-stimulations are sufficient to reinforce lever-

287 pressing behaviour. The stimulated hemisphere was counterbalanced within each group. After the

288 extinction sessions, one rat in the BLA-ChR2 group was excluded from subsequent testing because of

289 increasing aggressive behavior.

290

291 In this and subsequent experiments, the experimenter observed each rat during testing. Some

292 rats experienced seizures with repeated photo-stimulation of ChR2-containing BLA neurons (rats in

293 the other groups did not show seizure activity). This is consistent with the amygdala kindling model of

294 epilepsy and neuronal plasticity (Goddard et al., 1969; McNamara et al., 1980; Fisher, 1989). Rats that

295 experienced seizures were eliminated from final data analyses (Exp. 2, n = 0; Exp. 3a, n = 3; Exp. 3b,

296 n = 1), except for one rat in Exp. 3a (see below).

297

298 Exp. 3a: Effects of photo-stimulating ChR2-expressing BLA neurons during Pavlovian CS-UCS

299 conditioning on CS-evoked conditioned approach

300

301 Exp. 2 showed that rats reliably lever pressed for photo-stimulation of CeA but not BLA

302 neurons. Thus, we pursued the following experiments with BLA manipulations only, as the reinforcing

303 effects of CeA photo-stimulation could confound data interpretation. We first determined whether

304 photo-stimulation of BLA neurons during Pavlovian conditioning changes CS-evoked conditioned

305 approach behaviour, as measured by the CSR/PCSR ratio described above. As shown in Fig. 4A, a

306 new cohort of rats was prepared for optogenetic manipulations in the BLA as described above. The

307 rats then received Pavlovian conditioning under one of the following 3 conditions: 1) ‘No Laser’, where

308 the CS was presented alone (ChR2 = 11, eYFP = 5), 2) ‘Paired laser’, where photo-stimulation was

309 paired with each CS presentation (ChR2, n = 3; eYFP, n = 3), and 3) ‘Unpaired laser’, where photo-

310 stimulation and CS presentation were explicitly unpaired, by administering laser stimulation half-way

311 between each CS-UCS presentation (ChR2 = 3). The ‘Unpaired laser’ group served to determine

312 whether increased BLA neuronal activity had to coincide with CS presentation to influence CS-evoked

313 conditioned approach. If so, then CSR/PCSR ratios in the ‘Unpaired laser’ group should be similar to

314 those in the ’ChR2-No Laser’ or eYFP rats. One Unpaired-ChR2 rat had a seizure on session 9.

315 Therefore, the effects of BLA photo-stimulation on CSR/PCSR ratios were analysed on sessions 1-8,

316 with this rat included. There were no behavioural differences between ChR2-No laser, eYFP-Paired

317 laser and eYFP-No laser rats under any test condition, and they were pooled into one group (controls,

318 n = 19).

319

320 Exp. 3b: Effects of photo-stimulating ChR2-expressing BLA neurons during operant

321 responding for a CS

322

323 Rats naïve to laser stimulation (control rats from Exp. 3a, including 7 eYFP rats, and 8 ChR2

324 rats) received sessions where they could lever press for presentations of the CS, with or without CS-

325 paired BLA photo-stimulation (0, 5, 10 or 20 Hz, one frequency/session, counterbalanced), as shown

326 in Fig. 5A. We then determined whether photo-stimulation of BLA neurons must be explicitly paired

327 with CS presentations to alter operant responding for the CS. If so, then explicitly unpairing photo-

328 stimulation and CS presentation during operant responding for the CS should have no or reduced

329 effects on lever-pressing for that CS, compared to effects seen when photo-stimulation and CS

330 presentation are paired. To address this, all rats were given an operant responding session during

331 which photo-stimulation was explicitly unpaired with CS presentation (photo-stimulation applied 3

332 seconds after each CS presentation).

333

334 Exp. 4: Effects of intra-amygdala d-amphetamine infusions on the incentive motivational

335 effects of a CS

336

337 Exp. 3b showed that photo-stimulation of BLA neurons potentiates operant responding for a

338 CS, suggesting that changes in BLA neuron activity influences the incentive motivational effects of the

339 CS. Here we sought to extend these findings by using a pharmacological approach to influence BLA

340 neuron activity. Thus, we determined whether injecting d-amphetamine into the BLA (n = 20) also

341 changes operant responding for a CS. We also determined if, within the amygdala, effects of d-

342 amphetamine on CS incentive properties are specific to the BLA. To this end, we assessed the effects

343 of infusing d-amphetamine into the amygdala, but without targeting the BLA specifically (n = 15). We

344 predicted that d-amphetamine infused specifically into the BLA would enhance operant responding for

345 a CS, based on work showing that intra-BLA infusions of d-amphetamine increase cue-induced

346 reinstatement of extinguished cocaine seeking (Ledford et al., 2003). As shown in Fig. 7A, following

347 Pavlovian CS-UCS conditioning, intra-cerebral cannulae were implanted bilaterally. The rats were then

348 given at least 2 weeks to recover. Rats then received a reminder Pavlovian conditioning session. Right

349 after this session, rats received intra-cerebral saline infusions to habituate them to the infusion

350 procedure. No behavior was recorded. On the next day, rats received a lever habituation session.

351 Starting on the next day, rats received intra-cerebral saline or d-amphetamine (10 or 30

352 μg/hemisphere; Sigma-Aldrich, Dorset, UK; 1 injection/day, given every other day) and they were then

353 allowed to lever press for the CS during a 40-minute test session. This session length was chosen

354 based on our previous work with intra-nucleus accumbens d-amphetamine injections (El Hage et al.,

355 2015). Each rat received a maximum of 3 intra-cerebral injections to minimize tissue damage. This

356 included 1) a saline microinjection for habituation, 2) a saline microinjection prior to testing, and 3) a d-

357 amphetamine microinjection (10 or 30 μg/hemisphere) prior to testing (10 μg/hemisphere: n = 11 in

358 BLA group, n = 7 in Amygdala group; 30 μg/hemisphere: n = 9 in BLA group, n = 8 in Amygdala

359 group). Therefore, each rat received only one d-amphetamine microinjection. For intracerebral

360 injections, injectors (33 GA, model C315I, HRS Scientific) were inserted to extend 2 mm beyond the

361 cannulae. Microinjections were given in a volume of 0.5 μL/hemisphere and were infused over 1

362 minute using a microsyringe pump (HARVARD PHD, 2000: HARVARD Apparatus, Saint-Laurent,

363 Canada). Injectors were left in place for an additional minute after the infusion.

364

365 Histology

366

367 In Exps. 2-3, rats were anesthetized with urethane (1.2 g/kg, i.p.) and were transcardially

368 perfused with phosphate buffered saline and 4% paraformaldehyde. Brains were then extracted and

369 kept at room temperature for 1 week in a 30% sucrose/4% paraformaldehyde solution, and then

370 stored at -80 °C. In Exp. 4, rats were anesthetized with isoflurane (5 %), brains were extracted and

371 stored at -20 °C. Forty μm-thick coronal slices were cut in a cryostat and optic fiber or injector

372 placement was estimated using the Paxinos and Watson atlas (Paxinos and Watson, 1986).

373

374 Statistics

375

376 In Exp. 2, mixed-model ANOVA was used to analyse group differences in self-administered

377 photo-stimulations and lever pressing behaviour (Group × Session; ‘Session’ as a within-subjects

378 variable; Group x Time; ‘Time’ as a within-subjects variable; Group x Laser Frequency; ‘Frequency’ as

379 a within-subjects variable). One-way ANOVA was used to analyse group differences in both active

380 lever presses during the PR5 session and the number of self-administered unilateral stimulations. In

381 Exp. 3a, mixed-model ANOVA was used to analyse group differences in average CSR/PCSR ratios

382 (Group × Session; ‘Session’ as a within-subjects variable). In Exp. 3b, mixed-model ANOVA was used

383 to analyse group differences in lever pressing for the CS (Group × Session or Lever Type; ‘Session’

384 and ‘Lever Type’ as within-subjects variables). In Exp. 4, one-way ANOVA was used to analyse

385 CSR/PCSR ratios across sessions. The effects of d-amphetamine on lever pressing were analysed

386 using mixed-model ANOVA (Dose × Lever Type; ‘Lever Type’ as a within-subjects variable). When an

387 interaction and/or main effects were significant (p < 0.05), effects were analysed further using

388 Bonferroni’s multiple comparisons’ tests. Values in figures are mean ± SEM.

389

390 RESULTS

391

392 Exp. 1: Effects of photo-stimulation on action potentials in ChR2-expressing BLA and CeA

393 neurons in vivo

394

395 Figs. 2B-C show ChR2-eYFP expression in the BLA and CeA. As seen in Figs. 2D-E, photo-

396 stimulation of BLA or CeA neurons induced action potentials on average 100 % of the time at 1, 10

397 and 20 Hz stimulation frequencies. However, at 40 Hz, spike fidelity decreased, and photo-stimulation

398 produced action potentials only ~45 % of the time. In line with these observations, Figs. 2F-G show

399 that the frequencies of neuron firing and photo-stimulation were closely matched at laser frequencies ≤

400 20 Hz. However, at a stimulation frequency of 40 Hz, BLA and CeA neurons fired only at ~18 Hz. This

401 loss of fidelity is in accordance with the kinetic properties of ChR2(H134R), the ChR2 mutant used

402 here. Indeed, when 5-ms pulses are given at a 40-Hz stimulation frequency, pulses are spaced by 20

403 ms, and this is shorter than the combined opening (~3 ms) and closing (~18 ms) rates of

404 ChR2(H134R) (Lin et al., 2009). Importantly, laser application produced action potentials in ChR2-

405 expressing BLA or CeA neurons (Fig. 2H), but not in eYFP-expressing BLA or CeA neurons (Fig. 2I).

406 Thus, photo-stimulation reliably induced action potentials only in ChR2-expressing BLA or CeA

407 neurons, and spike fidelity was excellent at laser frequencies ≤ 20 Hz. Thus, we used frequencies ≤ 20

408 Hz in the following studies.

409

410 Exp. 2: Effects of photo-stimulation of BLA or CeA ChR2-expressing neurons on lever-pressing

411 behaviour

412

413 Here, we determined whether rats would reliably press on a lever for photo-stimulation of

414 ChR2-expressing BLA or CeA neurons (Fig. 3A). Pressing on the active lever produced photo-

415 stimulation paired with a light-tone cue, under FR1, RR2 and RR4 schedules of reinforcement (1

416 schedule/session). Pressing on the inactive lever had no programmed consequences.

417

418 Laser self-stimulation. Fig. 3B shows estimated optic fiber placements in the CeA and BLA.

419 Fig. 3C shows that across different reinforcement schedules, CeA-ChR2 rats self-administered more

420 laser stimulations than control rats (main effect of Group, F(2, 17) = 5.5, p = 0.014; CeA-ChR2 versus

421 control rats, F(1, 13) = 7.5, p = 0.017). Accordingly, as seen in Fig. 3D, CeA-ChR2 rats also pressed

422 more on the active lever than control rats (main effect of Group, F(2, 17) = 5.53, p = 0.014; CeA-ChR2

423 versus control rats, F(1, 13) = 7.42, p = 0.017). In contrast, BLA-ChR2 and control rats earned a

424 similar number of photo-stimulations and pressed a similar number of times on the active lever (Figs.

425 3C-D; all P’s > 0.05). Presses on the inactive lever did not differ between groups (Fig. 3E; p > 0.05),

426 suggesting that photo-stimulation of either BLA or CeA neurons did not produce nonspecific motor

427 effects. Fig. 3F shows the number of active lever presses for photo-stimulation under a PR5 schedule

428 of reinforcement. Under this schedule, CeA-ChR2 rats pressed more on the active lever relative to

429 BLA-ChR2 or control rats (main effect of Group, F(2, 17) = 3.77, p = 0.007; CeA-ChR2 > controls, p =

430 0.014; CeA-ChR2 > BLA-ChR2, p = 0.013). BLA-ChR2 rats and controls were not different (p > 0.05).

431 Thus, across a range of schedules of reinforcement, rats self-administered cued photo-stimulations of

432 CeA neurons, but not BLA neurons. The findings suggest that photo-stimulation of CeA, but not BLA

433 neurons is reinforcing.

434

435 Extinction responding. We assessed lever-pressing behaviour under extinction conditions

436 during a 40-minute session where photo-stimulation was only available from minutes 0-5 and 20-25,

437 under a RR2 schedule. As shown in Fig. 3G (top panel), BLA-ChR2 rats did not differ from controls

438 during this session (all P’s > 0.05). Presses on the inactive lever also did not differ between groups

439 (Fig. 3G, bottom panel; p > 0.05). However, Fig. 3G (top panel) also shows that when photo-

440 stimulation was available in the first 5 minutes of the session, CeA-ChR2 rats pressed more on the

441 active lever relative to controls and BLA-ChR2 rats (Group × Time interaction, F(14, 119) = 4.14, p <

442 0.0001; main effect of Group, F(2, 17) = 7.69, p = 0.004; CeA-ChR2 versus control rats, F(1, 13) =

443 10.3, p = 0.007; minutes 0-5, CeA-ChR2 > controls, p < 0.0001; CeA-ChR2 versus BLA-ChR2, F(1, 8)

444 = 5.11, p = 0.054; post-hoc comparisons on minutes 0-5, CeA-ChR2 > BLA-ChR2, p = 0.0001. No

445 other comparisons were significant). CeA-ChR2 rats also extinguished their lever-pressing behaviour

446 during the extinction session (Fig. 3G, top panel; main effect of Time, F(7, 119) = 6.12, p < 0.0001;

447 minutes 0-5 vs. each subsequent 5-minute block, all P’s < 0.0001). Thus, only CeA-ChR2 rats lever-

448 pressed for photo-stimulation when it was available, and decreased responding when it was not. In

449 contrast, BLA-ChR2 rats and control rats lever-pressed very little, regardless of photo-stimulation

450 availability.

451

452 Self-stimulation as a function of laser stimulation frequency. Fig. 3H shows the influence of

453 stimulation frequency (5, 10 and 20 Hz) on self-administration of photo-stimulations. Sessions stopped

454 after 30 stimulations or 30 minutes. As a measure of the rate of responding, we analysed the number

455 of photo-stimulations earned per minute. Relative to control rats, CeA-ChR2 rats earned more photo-

456 stimulations/minute at 10 and 20 Hz (Fig. 3H; Frequency × Group interaction, F(4, 32) = 4.08; p =

457 0.009; main effect of Group, F(2, 16) = 5.76, p = 0.013; CeA-ChR2 rats versus controls, F(1, 13) =

458 10.67, p = 0.006; CeA-ChR2 > controls at 10 Hz, p = 0.046, at 20 Hz, p < 0.0001). CeA-ChR2 rats

459 also earned more photo-stimulations/minute as stimulation frequency was increased (Fig. 3H; main

460 effect of Frequency, F(2, 32) = 9.31, p = 0.0006; CeA-ChR2 rats, 10 > 5 Hz, p = 0.019, 20 > 5 Hz, p <

461 0.0001). BLA-ChR2 rats earned more photo-stimulations/minute relative to controls only at the highest

462 frequency tested (main effect of Group, F(1, 12) = 6.18, p = 0.029; BLA-ChR2 > controls, at 20 Hz, p =

463 0.013). No other comparisons were statistically significant. Thus, compared to control rats, BLA-ChR2

464 rats earned more photo-stimulations/minute at 20 Hz, whereas CeA-ChR2 rats earned more photo-

465 stimulations/minute at both 10 and 20 Hz. Furthermore, only CeA-ChR2 rats increased their self-

466 stimulation behaviour with increasing laser frequency.

467

468 Reversal learning. Here we determined whether photo-stimulation of CeA or BLA neurons

469 supports reversal learning. Fig. 3I shows pressing on a lever that produced laser stimulation on

470 session ‘-1’, but not on subsequent sessions. Fig. 3J shows pressing on a lever that did not produce

471 laser stimulation on session ‘-1’, but did so on subsequent sessions. As seen in Fig. 3I, CeA-ChR2 but

472 not BLA-ChR2 rats pressed more on the reinforced lever relative to control rats (Group × Session

473 interaction, F(4, 32) = 5.46, p = 0.002; main effect of Group, F(2, 16) = 10.05, p = 0.002; CeA-ChR2

474 versus controls, F(1, 13) = 19.47, p = 0.0007; CeA-ChR2 > controls on session ‘-1’, p < 0.0001). CeA-

475 ChR2 rats also pressed significantly less on this lever after reversal versus before (main effect of

476 Session, F(2, 32) = 11.98, p = 0.0001; CeA-ChR2 rats, session -1 > session 1, p = 0.0002, session -1

477 > session 2, p < 0.0001). As seen in Fig. 3J, after lever reversal, CeA-ChR2 but not BLA-ChR2 rats

478 pressed more on the newly reinforced lever relative to controls (Group × Session interaction, F(4, 32)

479 = 5.94, p = 0.001; main effect of Group, F(2, 16) = 8.59, p = 0.003; CeA-ChR2 rats > controls, session

480 1, p = 0.033, session 2, p < 0.0001). CeA-ChR2 rats also pressed more on this lever after reversal

481 versus before (main effect of Session, F(2, 32) = 11.84, p = 0.0001; CeA-ChR2 rats, session -1 >

482 session 1, p = 0.035, session -1 > session 2, p < 0.0001). In summary, photo-stimulation of CeA

483 neurons both reliably reinforced lever-pressing behaviour and supported reversal learning, whereas

484 photo-stimulation of BLA neurons supported neither response.

485

486 Unilateral laser stimulation. Lastly, we determined whether unilateral photo-stimulation of CeA

487 or BLA neurons was reinforcing. Fig. 3K shows that CeA-ChR2 but not BLA-ChR2 rats earned more

488 unilateral laser stimulations relative to controls (main effect of Group, F(2, 16) = 6.24, p = 0.01; CeA-

489 ChR2 > controls, p = 0.009). Therefore, unilateral stimulation of CeA, but not BLA neurons sustains

490 self-stimulation.

491

492 In summary, across different schedules of reinforcement, operant testing conditions and photo-

493 stimulation parameters, rats did not reliably self-administer photo-stimulation of BLA neurons. In

494 contrast, rats reliably self-administered photo-stimulation of CeA neurons, indicating that it is

495 reinforcing. These findings show that photo-stimulation of BLA versus CeA neurons has dissociable

496 effects, and that CeA but not BLA neurons carry a primary reward signal.

497

498 Exp. 3a: Effects of photo-stimulating ChR2-expressing BLA neurons during Pavlovian CS-UCS

499 conditioning on CS-evoked conditioned approach

500

501 Fig. 4B shows estimated optic fiber placements in the BLA. We first determined the effects of

502 BLA photo-stimulation on CS-evoked conditioned approach behaviour (Fig. 4A). This was assessed

503 by analysing the ratio of nose-pokes into the water receptacle during each 5-second light cue

504 presentation (a conditioned stimulus response, or CSR), versus during the 5-second period preceding

505 each CS presentation (pre-conditioned stimulus response, or PCSR). Fig. 4C shows the effects of

506 photo-stimulation of ChR2-expressing BLA neurons on the CSR/PCSR ratio over Pavlovian

507 conditioning sessions. Average CSR/PCSR ratios progressively increased over sessions in all groups,

508 indicating that rats learned the CS-UCS contingency (Fig. 4C; main effect of Session F(3, 66) = 20.12,

509 p < 0.0001). Pairing photo-stimulation of BLA neurons with CS presentations (‘ChR2-Paired laser’

510 group) potentiated conditioned approach behaviour relative to all other conditions (Fig. 4C; Group ×

511 Session interaction, F(6, 66) = 8.31, p < 0.0001; main effect of Group, F(2, 22) = 21.81, p < 0.0001;

512 ChR2 Paired laser > Controls, session block 2, p = 0.025, session block 3, p < 0.0001, session block

513 4, p < 0.0001; ChR2 Paired laser > ChR2 Unpaired laser, session block 3, p < 0.0001, session block

514 4, p = 0.0004). No other comparisons were significant. Thus, photo-stimulation of BLA neurons

515 potentiated CS-evoked conditioned approach behaviour over time, but only if photo-stimulation was

516 explicitly paired with CS presentation.

517

518 Exp. 3b: Effects of photo-stimulating ChR2-expressing BLA neurons during operant

519 responding for a CS

520

521 Here, we sought to determine whether BLA photo-stimulation would potentiate instrumental

522 pursuit of the CS. To this end, we used rats from Exp. 3a that had undergone Pavlovian CS-UCS

523 conditioning without laser stimulation. These are rats with ChR2-expressing BLA neurons that had not

524 received laser photo-stimulation, and rats with eYFP-expressing BLA neurons. We determined in

525 these rats whether BLA photo-stimulation during operant responding for the CS enhances responding

526 for that CS (Fig. 5A). Fig. 5B shows presses on an active lever that produced CS presentation and on

527 an inactive lever, during a session where rats did not receive laser stimulation. Across groups, rats

528 pressed more on the active versus inactive lever (main effect of Lever Type, F(1, 13) = 13.86, p =

529 0.003). This indicates that the CS acquired incentive properties. There was neither a main effect of

530 Group nor a Group × Lever Type interaction effect (all P’s > 0.05). Thus, without laser stimulation,

531 ChR2 and eYFP rats show similar incentive motivation for the CS.

532

533 Fig. 5C shows presses on the active and inactive levers when CS presentations were paired

534 with BLA photo-stimulation at different laser frequencies (5, 10 or 20 Hz). Fig. 5C shows that both

535 ChR2 and eYFP rats pressed more on the active versus inactive lever (main effect of Lever Type, F(1,

536 26) = 18.3, p = 0.001; eYFP rats, F(1, 12) = 6.31, p = 0.027; ChR2 rats, F(1, 14) = 12.19, p = 0.004).

537 Thus, both groups showed incentive motivation for the CS under these conditions. In addition, ChR2

538 rats pressed more on the active lever than did eYFP rats (Fig. 5C; main effect of Group, F(1, 13) =

539 5.39, p = 0.04. No other comparisons were significant). This suggests that photo-stimulation of BLA

540 neurons potentiates the expression of incentive motivation for the CS. Fig. 5D shows lever-pressing

541 behaviour when pressing the active lever produced the CS and photo-stimulation 3 seconds later,

542 such that the CS and photo-stimulation were unpaired. Only ChR2 rats pressed more on the active

543 versus inactive lever (Fig. 5D; Group × Lever Type interaction, F(1, 13) = 5.79, p = 0.032; main effect

544 of Lever Type, F(1, 13) = 20.13, p = 0.0006; ChR2 rats, active > inactive lever, p = 0.0004). However,

545 ChR2 rats did not press more on the active lever than eYFP rats (p > 0.05). No other comparisons

546 were significant. Lastly, BLA photo-stimulation, when it was either paired or unpaired with CS

547 presentation, did not influence nose pokes into the water receptacle (data not shown, all P’s > 0.05).

548 This suggests that BLA photo-stimulation did not increase the urge to consume the associated water

549 UCS. Taken together, these results show that photo-stimulation of BLA neurons during operant

550 responding for the CS potentiated incentive motivation for that CS, and that this effect was strongest

551 when photo-stimulation was explicitly paired with each CS presentation.

552

553 Together, the results of Exp. 2 and 3b indicate that BLA photo-stimulation increases

554 instrumental pursuit of a discrete stimulus, if and only if that stimulus reliably predicts a primary reward

555 (water). That is, BLA photo-stimulation selectively potentiates the pursuit of environmental stimuli that

556 possess conditioned incentive properties. Fig. 6 highlights this effect. It shows reinforcements earned

557 by individual rats lever pressing for presentations of a light-tone stimulus not previously associated

558 with a reward (Figs. 6A-B; rats from Exp. 2) or a light-tone stimulus previously associated with a water

559 reward (Figs. 6C-D; rats from Exp. 3b). When the light-tone stimulus had no relationship with a

560 primary reward, BLA photo-stimulation did not significantly change the number of stimulus

561 presentations earned (Figs. 6A-B). After the light-tone stimulus had been paired with water, control

562 rats pursued this CS more avidly (compare Figs. 6A and C), and BLA photo-stimulation potentiated

563 this effect (compare Figs. 6C and D).

564

565 Additionally, photo-stimulation of BLA neurons increased both CS-evoked conditioned

566 approach (Fig. 4C; Exp. 3a) and operant responding for the CS (Fig. 5C; Exp. 3b). These two CS

567 effects rely on common but also partially dissociable neurobiological and psychological processes

568 (Flagel et al., 2011; Tabbara et al., 2016). In agreement, in the control rats represented in Fig. 4C,

569 there was no significant correlation between average CSR/PCSR ratios over the last 4 days of

570 Pavlovian conditioning and active lever presses during a subsequent instrumental conditioning

571 session (without laser; data not shown; r2 = 0.002, p = 0.85).

572

573 Exp. 4: Effects of intra-amygdala d-amphetamine infusions on the incentive motivational

574 effects of a CS

575

576 After CS-UCS Pavlovian conditioning, rats were given instrumental responding tests where

577 they could lever-press for the CS (Fig. 7A). Immediately prior to these tests, rats received bilateral

578 infusions of d-amphetamine (0, 10 or 30 μg/hemisphere) into the BLA or into the amygdala without

579 targeting the BLA specifically. Fig. 7B shows estimated location of injector tips when both cannulae

580 were specifically in the BLA (top) or simply in the amygdala, but without targeting the BLA exclusively

581 (bottom). The rats learned the CS-UCS contingency, as indicated by a progressive increase in

582 CSR/PCSR ratio (Fig. 7C; main effect of Session, F(4, 76) = 11.12, p < 0.0001; Fig. 7F; main effect of

583 Session, F(4, 56) = 5.04, p = 0.002). Figs. 7D-E-G-H show that rats in both experimental groups

584 pressed more on the active versus inactive lever (Figs. 7D-E; Dose × Lever Type interaction, F(2, 37)

585 = 5.31, p = 0.009; main effect of Lever Type, F(1, 37) = 142.4; p < 0.0001; Figs. 7G-H; main effect of

586 Lever Type, F(1, 27) = 25.61, p < 0.0001). Thus, all rats spontaneously learned a new operant

587 response to produce the CS, indicating that the CS acquired incentive value. D-amphetamine

588 influenced active lever pressing only when infused specifically into the BLA, such that active lever

589 pressing was greatest at 30 μg/hemisphere d-amphetamine (Fig. 7D; main effect of Dose, F(2, 37) =

590 4.5, p = 0.018; 30 vs 0 μg, p = 0.0002; 30 vs 10 μg, p = 0.027). In contrast, d-amphetamine did not

591 alter lever-pressing behaviour in rats that received infusions into the amygdala, without specifically

592 targeting the BLA (Figs. 7G-H; all P's > 0.05). No other comparisons were statistically significant.

593 Lastly, neither intra-BLA nor intra-amygdala d-amphetamine altered the number of nose pokes into the

594 water receptacle (data not shown, all P’s > 0.05). This suggests that d-amphetamine infusions into the

595 amygdala did not increase the urge to consume the associated water UCS. Thus, the findings show

596 that intra-BLA d-amphetamine intensified incentive motivation for the CS.

597

598 DISCUSSION

599

600 We evaluated the contributions of the BLA to appetitive Pavlovian conditioning and to the

601 instrumental pursuit of a reward-predictive CS. First, photo-stimulation of BLA neurons was not

602 intrinsically reinforcing, whereas photo-stimulation of neurons in the adjacent CeA was. Second,

603 photo-stimulation of BLA neurons during Pavlovian CS-UCS conditioning enhanced CS-evoked

604 conditioned approach, indicating potentiated anticipation of the primary reward. Third, photo-

605 stimulation of BLA neurons potentiated operant responding for the CS, suggesting enhanced CS

606 incentive value. Finally, intra-BLA infusions of d-amphetamine also augmented operant responding for

607 the CS, suggesting that a local increase in monoamine neurotransmission is also involved in

608 enhanced conditioned incentive motivation. Thus, increased neuronal activity within the BLA facilitates

609 cue-controlled behaviour by both increasing cue-induced anticipation of impending rewards and

610 making reward cues more attractive.

611

612 Photo-stimulation of CeA, but not BLA neurons is reinforcing

613

614 Rats reliably lever pressed for photo-stimulation of CeA, but not BLA neurons, suggesting that

615 CeA neurons carry a primary reward signal. Our findings agree with earlier work showing that

616 electrical stimulation of CeA cells is reinforcing (Prado-Alcala and Wise, 1984; Kane et al., 1991). CeA

617 neurons are mostly GABAergic, but they express different neuropeptides and have different

618 anatomical connections. More recent studies show that stimulation of specific neuronal populations in

619 the CeA can also be reinforcing. This includes CeA neurons expressing corticotropin-releasing

620 hormone, somatostatin, neurotensin and/or tachykinin 2 (Baumgartner et al., 2017; Kim et al., 2017),

621 and CeAÆmedial prefrontal cortex neurons (Seo et al., 2016). In contrast, using photo-stimulation of

622 CeA neurons without regards to cell subtype as done here, Berridge and colleagues report that CeA

623 photo-stimulation is not reinforcing (Robinson et al., 2014; Warlow et al., 2017). This could involve the

624 CeA subregion where photo-stimulation was applied. Berridge and colleagues (Robinson et al., 2014;

625 Warlow et al., 2017) implanted optic fibers in the posterior CeA, whereas we implanted in the anterior

626 CeA. Our rats did not reliably self-administer photo-stimulation of BLA neurons. Rats will electrically

627 self-stimulate some BLA subregions (Prado-Alcala and Wise, 1984; Kane et al., 1991), and studies

628 using optogenetic methods suggest that self-stimulation depends on the BLA circuit targeted. For

629 instance, photo-stimulation of BLAÆnucleus accumbens terminals is reinforcing (Stuber et al., 2011;

630 Britt et al., 2012; Namburi et al., 2015), but photo-stimulation of BLAÆmedial CeA terminals produces

631 avoidance (Namburi et al., 2015). The absence of BLA self-stimulation here could involve the hSyn

632 promoter we used. It confers neuron-specific transgene expression, but it does not target neuron

633 subtypes.

634

635 Via distinct cell types and connections, amygdala nuclei and subregions exert many functions,

636 including both appetitive and defensive behaviours (Gallagher and Chiba, 1996). Future studies will be

637 important to examine roles of specific CeA and BLA neuron subtypes and projections in appetitive

638 behaviour. As this research unfolds, our results support the idea that while the BLA and CeA are

639 connected and can play similar roles in motivational processes (Wassum et al., 2011), they also have

640 distinct appetitive functions (Corbit and Balleine, 2005; Robinson et al., 2014; Warlow et al., 2017).

641

642 Photo-stimulation of BLA neurons during CS-UCS conditioning enhances CS-evoked

643 conditioned approach

644

645 During Pavlovian conditioning, we paired photo-stimulation of BLA neurons with CS

646 presentation. This potentiated CS-evoked conditioned approach, as shown by more CS-triggered

647 visits to the water dish. This suggests enhanced anticipation of the CS-associated water reward.

648 Explicitly unpairing BLA stimulation and CS presentation did not influence CS-evoked conditioned

649 approach. Thus, increasing BLA neuron activity when a CS is presented amplifies associative CS-

650 UCS learning. Increased CS-triggered visits to the water dish could suggest that BLA photo-

651 stimulation enhances the appetitive value of water. This is possible, but unlikely, because BLA lesions

652 do not alter water consumption (Cador et al., 1989). Instead, enhanced CS-evoked conditioned

653 approach likely involves changes in how BLA neurons represent the CS and/or how they encode the

654 CS-UCS association. CS-triggered conditioned approach behaviours can reflect both the predictive

655 and incentive effects of CS. BLA stimulation could increase CS-triggered visits to the water dish by

656 enhancing either or both effects. For instance, rats might visit the water dish during CS presentation

657 because the CS is evoking an incentive urge to drink the associated water (Weingarten, 1983). If so,

658 then BLA photo-stimulation during the CS could increase visits to the water dish by enhancing this

659 conditioned incentive urge [see for example Holland et al. (2002)]. Similarly, the water dish is also a

660 CS in our experiments, less predictive than the light-tone CS, but more proximal to the water UCS. As

661 such, BLA photo-stimulation could have increased water dish visits by enhancing the incentive value

662 of the dish. We do not believe this is the case, because BLA photo-stimulation increased the number

663 of water dish visits only when this stimulation was explicitly paired with the light-tone CS.

664

665 Photo-stimulation of BLA neurons or d-amphetamine infusion into the BLA enhances CS

666 incentive value

667

668 Once the CS had been imbued with incentive value through prior association with an appetitive

669 UCS, BLA photo-stimulation amplified the expression of this incentive motivation (as measured by

670 lever-pressing reinforced by the CS alone). Infusing d-amphetamine into the BLA had the same effect,

671 suggesting that increases in monoamine-mediated neurotransmission in the BLA are involved

672 (Ledford et al., 2003; Bernardi et al., 2009; Gremel and Cunningham, 2009; Lintas et al., 2011). This

673 extends lesion studies showing that the BLA is necessary for operant responding reinforced by a CS

674 (Cador et al., 1989; Burns et al., 1993). BLA photostimulation or d-amphetamine infusions into the

675 BLA could have enhanced instrumental responding for the CS by potentiating the appetitive value of

676 the associated water reward. This is unlikely, because neither manipulation influenced the number of

677 water dish visits during instrumental tests. In addition, the increased lever pressing during tests of

678 instrumental responding for the CS likely does not involve any intrinsically reinforcing effects of BLA

679 photo-stimulation. Indeed, our BLA photo-stimulation parameters did not reliably support self-

680 stimulation behaviour. Instead, the BLA stores information about CS value, which is then used to

681 guide behaviour (Cardinal et al., 2002). As such, stimulation of BLA neurons could enhance operant

682 responding for a CS by potentiating the incentive value of the CS itself or of the CS-associated reward

683 representation (Mogenson, 1987; Everitt and Robbins, 1992).

684

685 Conclusions

686

687 Increased neuronal activity in BLA-dependent circuits amplifies control over behaviour by an

688 appetitive cue, and this involves two overlapping, but also dissociable psychological mechanisms. A

689 first mechanism involves enhanced CS-UCS associative learning, such that the CS triggers increased

690 conditioned approach, and increased anticipation of the primary reward. This prepares animals to

691 engage with the forthcoming reward. The second mechanism involves amplified incentive motivation

692 to pursue the CS, such that animals show enhanced instrumental responding for the CS. Thus, when

693 reward cues are present in the environment, increased recruitment of BLA-dependent pathways could

694 promote excessive pursuit of associated rewards both by augmenting anticipation for these rewards

695 and making reward-paired cues more attractive in their own right.

696

697 REFERENCES

698

699 Ambroggi F, Ishikawa A, Fields HL, Nicola SM (2008) Basolateral amygdala neurons facilitate reward-

700 seeking behavior by exciting nucleus accumbens neurons. Neuron 59:648-661.

701 Baumgartner HM, Olney JJ, Warlow SM, Schulkin J, Berridge KC (2017) Investigating corticotropin

702 releasing factor in mediating appetitive behavior. Published abstract for the Society for

703 Neuroscience Annual Meeting, Washington, DC.

704 Bernardi RE, Ryabinin AE, Berger SP, Lattal KM (2009) Post-retrieval disruption of a cocaine

705 conditioned place preference by systemic and intrabasolateral amygdala beta2- and alpha1-

706 adrenergic antagonists. Learn Mem 16:777-789.

707 Bindra D (1978) How adaptive behavior is produced: a perceptual-motivation alternative to response

708 reinforcement. Behavioral and Brain Sciences 1:41-52.

709 Bolles RC (1972) Reinforcement, expectancy, and learning. Psychological Review 79:394-409.

710 Britt JP, Benaliouad F, McDevitt RA, Stuber GD, Wise RA, Bonci A (2012) Synaptic and behavioral

711 profile of multiple glutamatergic inputs to the nucleus accumbens. Neuron 76:790-803.

712 Brown EE, Fibiger HC (1993) Differential effects of excitotoxic lesions of the amygdala on cocaine-

713 induced conditioned locomotion and conditioned place preference. Psychopharmacology

714 113:123-130.

715 Burns LH, Robbins TW, Everitt BJ (1993) Differential effects of excitotoxic lesions of the basolateral

716 amygdala, ventral subiculum and medial prefrontal cortex on responding with conditioned

717 reinforcement and locomotor activity potentiated by intra-accumbens infusions of D-

718 amphetamine. Behavioural brain research 55:167-183.

719 Cador M, Robbins TW, Everitt BJ (1989) Involvement of the amygdala in stimulus-reward

720 associations: interaction with the ventral striatum. Neuroscience 30:77-86.

721 Cardinal RN, Parkinson JA, Hall J, Everitt BJ (2002) Emotion and motivation: the role of the amygdala,

722 ventral striatum, and prefrontal cortex. Neurosci Biobehav Rev 26:321-352.

723 Corbit LH, Balleine BW (2005) Double dissociation of basolateral and central amygdala lesions on the

724 general and outcome-specific forms of pavlovian-instrumental transfer. The Journal of

725 neuroscience : the official journal of the Society for Neuroscience 25:962-970.

726 de Wit H, Stewart J (1981) Reinstatement of cocaine-reinforced responding in the rat.

727 Psychopharmacology 75:134-143.

728 Dittgen T, Nimmerjahn A, Komai S, Licznerski P, Waters J, Margrie TW, Helmchen F, Denk W, Brecht

729 M, Osten P (2004) Lentivirus-based genetic manipulations of cortical neurons and their optical

730 and electrophysiological monitoring in vivo. Proceedings of the National Academy of Sciences

731 of the United States of America 101:18206-18211.

732 El Hage C, Bedard AM, Samaha AN (2015) Antipsychotic treatment leading to dopamine

733 supersensitivity persistently alters nucleus accumbens function. Neuropharmacology.

734 Everitt B, Robbins T (1992) Amygdala-Ventral Striatal Interactions and Reward-related Processes.

735 The amygdala: Neurobiological aspects of emotion, memory, and mental dysfunction New

736 York, NY, US: Wiley-Liss.:401-409.

737 Everitt BJ, Morris KA, O'Brien A, Robbins TW (1991) The basolateral amygdala-ventral striatal system

738 and conditioned place preference: further evidence of limbic-striatal interactions underlying

739 reward-related processes. Neuroscience 42:1-18.

740 Fisher RS (1989) Animal models of the epilepsies. Brain Res Brain Res Rev 14:245-278.

741 Flagel SB, Akil H, Robinson TE (2009) Individual differences in the attribution of incentive salience to

742 reward-related cues: Implications for addiction. Neuropharmacology 56 Suppl 1:139-148.

743 Flagel SB, Clark JJ, Robinson TE, Mayo L, Czuj A, Willuhn I, Akers CA, Clinton SM, Phillips PE, Akil H

744 (2011) A selective role for dopamine in stimulus-reward learning. Nature 469:53-57.

745 Gabriele A, See RE (2010) Reversible inactivation of the basolateral amygdala, but not the

746 dorsolateral caudate putamen, attenuates consolidation of cocaine-cue associative learning in

747 a reinstatement model of drug-seeking. The European journal of neuroscience 32:1024-1029.

748 Gallagher M, Chiba AA (1996) The amygdala and emotion. Current opinion in neurobiology 6:221-227.

749 Goddard GV, McIntyre DC, Leech CK (1969) A permanent change in brain function resulting from

750 daily electrical stimulation. Exp Neurol 25:295-330.

751 Gremel CM, Cunningham CL (2009) Involvement of amygdala dopamine and nucleus accumbens

752 NMDA receptors in ethanol-seeking behavior in mice. Neuropsychopharmacology : official

753 publication of the American College of Neuropsychopharmacology 34:1443-1453.

754 Grimm JW, See RE (2000) Dissociation of primary and secondary reward-relevant limbic nuclei in an

755 animal model of relapse. Neuropsychopharmacology : official publication of the American

756 College of Neuropsychopharmacology 22:473-479.

757 Hearst E, Jenkins HM (1974) Sign-tracking: The Stimulus-reinforcer Relation and Directed Action.

758 Psychonomic Society, Austin.

759 Holland PC, Petrovich GD, Gallagher M (2002) The effects of amygdala lesions on conditioned

760 stimulus-potentiated eating in rats. Physiol Behav 76:117-129.

761 Huff ML, Miller RL, Deisseroth K, Moorman DE, LaLumiere RT (2013) Posttraining optogenetic

762 manipulations of basolateral amygdala activity modulate consolidation of inhibitory avoidance

763 memory in rats. Proceedings of the National Academy of Sciences of the United States of

764 America 110:3597-3602.

765 Ilango A, Kesner AJ, Broker CJ, Wang DV, Ikemoto S (2014) Phasic excitation of ventral tegmental

766 dopamine neurons potentiates the initiation of conditioned approach behavior: parametric and

767 reinforcement-schedule analyses. Front Behav Neurosci 8:155.

768 Kane F, Coulombe D, Miliaressis E (1991) Amygdaloid self-stimulation: a movable electrode mapping

769 study. Behav Neurosci 105:926-932.

770 Kantak KM, Black Y, Valencia E, Green-Jordan K, Eichenbaum HB (2002) Dissociable effects of

771 lidocaine inactivation of the rostral and caudal basolateral amygdala on the maintenance and

772 reinstatement of cocaine-seeking behavior in rats. The Journal of neuroscience : the official

773 journal of the Society for Neuroscience 22:1126-1136.

774 Kim J, Pignatelli M, Xu S, Itohara S, Tonegawa S (2016) Antagonistic negative and positive neurons of

775 the basolateral amygdala. Nat Neurosci 19:1636-1646.

776 Kim J, Zhang X, Muralidhar S, LeBlanc SA, Tonegawa S (2017) Basolateral to Central Amygdala

777 Neural Circuits for Appetitive Behaviors. Neuron 93:1464-1479 e1465.

778 Lajoie J, Bindra D (1976) An interpretation of autoshaping and related phenomena in terms of

779 stimulus-incentive contingencies alone. Can J Psychol 30:157-173.

780 Ledford CC, Fuchs RA, See RE (2003) Potentiated reinstatement of cocaine-seeking behavior

781 following D-amphetamine infusion into the basolateral amygdala. Neuropsychopharmacology :

782 official publication of the American College of Neuropsychopharmacology 28:1721-1729.

783 Lin JY, Lin MZ, Steinbach P, Tsien RY (2009) Characterization of engineered channelrhodopsin

784 variants with improved properties and kinetics. Biophys J 96:1803-1814.

785 Lintas A, Chi N, Lauzon NM, Bishop SF, Gholizadeh S, Sun N, Tan H, Laviolette SR (2011)

786 Identification of a dopamine receptor-mediated opiate reward memory switch in the basolateral

787 amygdala-nucleus accumbens circuit. The Journal of neuroscience : the official journal of the

788 Society for Neuroscience 31:11172-11183.

789 Mackintosh NJ (1974) The psychology of animal learning. London ; New York: Academic Press.

790 McDonald RJ, Hong NS (2004) A dissociation of dorso-lateral striatum and amygdala function on the

791 same stimulus-response habit task. Neuroscience 124:507-513.

792 McDonald RJ, Yim TT, Lehmann H, Sparks FT, Zelinski EL, Sutherland RJ, Hong NS (2010)

793 Expression of a conditioned place preference or spatial navigation task following muscimol-

794 induced inactivations of the amygdala or dorsal hippocampus: A double dissociation in the

795 retrograde direction. Brain research bulletin 83:29-37.

796 McLaughlin J, See RE (2003) Selective inactivation of the dorsomedial prefrontal cortex and the

797 basolateral amygdala attenuates conditioned-cued reinstatement of extinguished cocaine-

798 seeking behavior in rats. Psychopharmacology 168:57-65.

799 McNamara JO, Byrne MC, Dasheiff RM, Fitz JG (1980) The kindling model of epilepsy: a review.

800 Progress in neurobiology 15:139-159.

801 Millan EZ, Kim HA, Janak PH (2017) Optogenetic activation of amygdala projections to nucleus

802 accumbens can arrest conditioned and unconditioned alcohol consummatory behavior.

803 Neuroscience 360:106-117.

804 Mogenson GJ (1987) Limbic-motor integration. Prog Psychobiol Physiol Psvchol 12:117-170.

805 Namburi P, Beyeler A, Yorozu S, Calhoon GG, Halbert SA, Wichmann R, Holden SS, Mertens KL,

806 Anahtar M, Felix-Ortiz AC, Wickersham IR, Gray JM, Tye KM (2015) A circuit mechanism for

807 differentiating positive and negative associations. Nature 520:675-678.

808 Paxinos G, Watson C (1986) The rat brain in stereotaxic coordinates, 2nd Edition. Sydney ; Orlando:

809 Academic Press.

810 Prado-Alcala R, Wise RA (1984) Brain stimulation reward and dopamine terminal fields. I. Caudate-

811 putamen, nucleus accumbens and amygdala. Brain research 297:265-273.

812 Rescorla RA (1988) Pavlovian conditioning. It's not what you think it is. Am Psychol 43:151-160.

813 Rescorla RA, Solomon RL (1967) Two-process learning theory: Relationships between Pavlovian

814 conditioning and instrumental learning. Psychol Rev 74:151-182.

815 Robbins TW (1978) The acquisition of responding with conditioned reinforcement: effects of pipradrol,

816 methylphenidate, d-amphetamine, and nomifensine. Psychopharmacology 58:79-87.

817 Robinson MJ, Warlow SM, Berridge KC (2014) Optogenetic excitation of central amygdala amplifies

818 and narrows incentive motivation to pursue one reward above another. The Journal of

819 neuroscience : the official journal of the Society for Neuroscience 34:16567-16580.

820 Rogers JL, Ghee S, See RE (2008) The neural circuitry underlying reinstatement of heroin-seeking

821 behavior in an animal model of relapse. Neuroscience 151:579-588.

822 Rossi MA, Sukharnikova T, Hayrapetyan VY, Yang L, Yin HH (2013) Operant self-stimulation of

823 dopamine neurons in the substantia nigra. PLoS One 8:e65799.

824 Seo DO, Funderburk SC, Bhatti DL, Motard LE, Newbold D, Girven KS, McCall JG, Krashes M, Sparta

825 DR, Bruchas MR (2016) A GABAergic Projection from the Centromedial Nuclei of the

826 Amygdala to Ventromedial Prefrontal Cortex Modulates Reward Behavior. The Journal of

827 neuroscience : the official journal of the Society for Neuroscience 36:10831-10842.

828 Stefanik MT, Kalivas PW (2013) Optogenetic dissection of basolateral amygdala projections during

829 cue-induced reinstatement of cocaine seeking. Front Behav Neurosci 7:213.

830 Stuber GD, Sparta DR, Stamatakis AM, van Leeuwen WA, Hardjoprajitno JE, Cho S, Tye KM,

831 Kempadoo KA, Zhang F, Deisseroth K, Bonci A (2011) Excitatory transmission from the

832 amygdala to nucleus accumbens facilitates reward seeking. Nature 475:377-380.

833 Tabbara RI, Maddux JM, Beharry PF, Iannuzzi J, Chaudhri N (2016) Effects of sucrose concentration

834 and water deprivation on Pavlovian conditioning and responding for conditioned reinforcement.

835 Behav Neurosci 130:231-242.

836 Tolman EC (1932) Purposive behavior in animals and men. London, England: Century/Random

837 House UK.

838 Trujillo-Pisanty I, Sanio C, Chaudhri N, Shizgal P (2015) Robust optical fiber patch-cords for in vivo

839 optogenetic experiments in rats. MethodsX 2:263-271.

840 Tye KM, Janak PH (2007) Amygdala neurons differentially encode motivation and reinforcement. The

841 Journal of neuroscience : the official journal of the Society for Neuroscience 27:3937-3945.

842 Tye KM, Stuber GD, de Ridder B, Bonci A, Janak PH (2008) Rapid strengthening of thalamo-

843 amygdala synapses mediates cue-reward learning. Nature 453:1253-1257.

844 Warlow SM, Robinson MJF, Berridge KC (2017) Optogenetic Central Amygdala Stimulation Intensifies

845 and Narrows Motivation for Cocaine. The Journal of neuroscience : the official journal of the

846 Society for Neuroscience 37:8330-8348.

847 Wassum KM, Cely IC, Balleine BW, Maidment NT (2011) Micro-opioid receptor activation in the

848 basolateral amygdala mediates the learning of increases but not decreases in the incentive

849 value of a food reward. The Journal of neuroscience : the official journal of the Society for

850 Neuroscience 31:1591-1599.

851 Weingarten HP (1983) Conditioned cues elicit feeding in sated rats: a role for learning in meal

852 initiation. Science 220:431-433.

853 White NM, McDonald RJ (1993) Acquisition of a spatial conditioned place preference is impaired by

854 amygdala lesions and improved by fornix lesions. Behavioural brain research 55:269-281.

855 Zhang F, Gradinaru V, Adamantidis AR, Durand R, Airan RD, de Lecea L, Deisseroth K (2010)

856 Optogenetic interrogation of neural circuits: technology for probing mammalian brain

857 structures. Nature protocols 5:439-456.

858

859 FIGURE LEGENDS

860

861 Figure 1. Pavlovian and Instrumental conditioning procedures. (A) During Pavlovian conditioning,

862 rats with limited access to water (2 h/day) learned that a cue (lights + tone, CS) predicts water (100 μl)

863 delivery into a recessed dish. We assessed the acquisition of CS-evoked conditioned approach

864 behaviour by analysing the ratio of the number of nose-pokes into the dish made during each 5-

865 second cue presentation (conditioned stimulus response; CSR) over that made during the 5 seconds

866 prior to each CS presentation (pre-conditioned stimulus response; PCSR). (B) After Pavlovian

867 conditioning, rats were given instrumental conditioning sessions during which they were presented

868 with two levers for the first time. Pressing the active lever produced the CS, while pressing the inactive

869 lever had no programmed outcome. No water was delivered during instrumental conditioning

870 sessions.

871

872 Figure 2. Photo-stimulation reliably induces action potentials only in basolateral (BLA) and

873 central (CeA) amygdala neurons expressing ChR2. (A) In Exp. 1, rats received AAV5-hSyn1-

874 hChR2(H134R)-eYFP (ChR2-eYFP) for transduction and activation of BLA or CeA neurons. Control

875 rats received an optically inactive virus lacking ChR2 (AAV5-hSyn1-eYFP) in the BLA or CeA. At least

876 4 weeks later, we measured action potentials evoked by photo-stimulation using in vivo

877 electrophysiology. (B, C) show ChR2-eYFP expression in the BLA and CeA, respectively (scale bars:

878 50 μm, arrows indicate cell bodies). When laser-light is delivered, ChR2 reliably induced action

879 potentials in (D) BLA and (E) CeA neurons, with stimulation frequencies ranging between 1-20 Hz.

880 ChR2 fidelity was reduced at 40 Hz. Accordingly, firing frequency of (F) BLA and (G) CeA neurons

881 matched laser stimulation frequency only between 1-20 Hz. Recordings in 4 rats/region; 10

882 observations/rat. Examples of in vivo recordings show that laser-light induced action potentials in (H)

883 ChR2-expressing BLA and CeA neurons but not in (I) eYFP-expressing BLA and CeA neurons.

884

885 Figure 3. Photo-stimulation of neurons in the central (CeA), but not basolateral (BLA) amygdala

886 is reinforcing. (A) In Exp. 2, rats received AAV5-hSyn1-hChR2(H134R)-eYFP or an optically inactive

887 control virus lacking ChR2 (AAV5-hSyn1-eYFP) in the BLA or CeA of both hemispheres. Optic fibers

888 were also implanted bilaterally, above virus injection sites. (B) Estimated optic fiber placements in the

889 CeA and BLA (anteroposterior position is shown in mm relative to Bregma). At least 4 weeks after

890 surgery, rats were allowed to press on two levers. Pressing the active lever produced photo-

891 stimulation of BLA or CeA neurons, paired with presentation of a light-tone cue. Pressing the inactive

892 lever had no programmed consequence. (C-E) Lever pressing was measured under fixed ratio 1

893 (FR1), random ratio 2 (RR2) and random ratio 4 (RR4) schedules of laser reinforcement. Responding

894 was also assessed under (F) a progressive ratio 5 (PR5) schedule of laser reinforcement, and during

895 a (G) within-session extinction test. (H) Effects of laser stimulation frequency on stimulations

896 earned/minute. (I, J) Lever pressing under reversal learning conditions. (K) Effects of unilateral

897 stimulation, under a RR2 schedule of laser reinforcement. *p < 0.05. In (G), # p < 0.05, versus control

898 rats and BLA-ChR2 rats; α p < 0.05, 1st 5-minute block versus all other 5-minute blocks in CeA-ChR2

899 rats. In (H), # p < 0.05, versus control rats at the same frequency; α p < 0.05, versus 5 Hz in CeA-

900 ChR2 rats. In (I), # p < 0.05, versus control rats in session -1; α p < 0.05, versus sessions 1 and 2 in

901 CeA-ChR2 rats. In (J), # p < 0.05, versus control rats in the same test session; α p < 0.05, versus

902 session -1 in CeA-ChR2 rats. n’s = 4-10/group.

903

904 Figure 4. Photo-stimulation of basolateral amygdala (BLA) neurons during conditioned

905 stimulus (CS) presentation potentiates CS-evoked conditioned approach. (A) In Exp. 3a, rats

906 received AAV5-hSyn1-hChR2(H134R)-eYFP or an optically inactive control virus lacking ChR2

907 (AAV5-hSyn1-eYFP) in the BLA of both hemispheres. Optic fibers were also implanted bilaterally,

908 above virus injection sites. (B) Estimated optic fiber placements in the BLA (anteroposterior position is

909 shown in mm relative to Bregma). At least 4 weeks after surgery, rats were water-restricted (2 h/day)

910 and received Pavlovian conditioning sessions where a light-tone conditioned stimulus (CS) predicted

911 water (100 μl) delivery (unconditioned stimulus, or UCS) into a recessed dish. During conditioning

912 sessions, photo-stimulation of BLA neurons was either explicitly paired or unpaired with CS

913 presentation, in independent groups of rats. Control rats included ChR2 and eYFP rats that did not

914 receive photo-stimulations and eYFP rats that received photo-stimulations. (C) CS-paired but not CS-

915 unpaired BLA photo-stimulation enhanced CSR/PCSR ratios (ratio of nose-pokes into the water dish

916 during each 5-second CS presentation versus nose-pokes made during the 5-second period

917 preceding each CS presentation). This indicates enhanced Pavlovian learning. n’s = 3-19/group. *p <

918 0.05, versus ChR2-Unpaired laser group and control group; # p < 0.05, versus control group on the

919 same session; α p < 0.05, versus ChR2-Unpaired laser group on the same session.

920

921 Figure 5. Photo-stimulation of basolateral amygdala (BLA) neurons potentiates incentive

922 motivation for a conditioned stimulus (CS). (A) In Exp. 3b, rats that had not received photo-

923 stimulation of BLA neurons during previous Pavlovian CS-UCS conditioning (eYFP rats and ChR2-No

924 laser control rats from Exp. 3a) were used to assess the effects of photo-stimulation of BLA neurons

925 during instrumental responding for the CS. (B) During a session without laser stimulation, both groups

926 pressed more on the active versus inactive lever, and there were no group differences in lever-

927 pressing behaviour. (C) During sessions where BLA photo-stimulation was paired with each earned

928 CS presentation, ChR2 rats pressed more on the active lever than eYFP rats did. This indicates that

929 photo-stimulation of BLA neurons during CS presentation enhances the incentive motivational value of

930 the CS. (D) During sessions where BLA photo-stimulation was explicitly unpaired with each earned CS

931 presentation, ChR2 rats still pressed more on the active versus inactive lever, but lever-pressing

932 behavior did not differ between ChR2 and eYFP rats. n’s =7-8/ group. *p < 0.05; α p < 0.05, active

933 lever presses versus inactive lever presses in ChR2 rats.

934

935 Figure 6. Photo-stimulation of basolateral amygdala (BLA) neurons potentiates motivation for a

936 discrete environmental stimulus if and only if this stimulus was previously associated with a

937 primary reward. Each dot indicates reinforcements earned by individual rats that were lever pressing

938 for presentations of a light-tone stimulus. Data is shown for individual control rats (A and C) and

939 individual rats receiving photo-stimulation of ChR2-expressing BLA neurons paired with each stimulus

940 presentation (B and D). (A-B) When the light-tone stimulus had not previously been associated with a

941 primary reward, BLA photo-stimulation did not change the number of stimulus presentations earned.

942 (C-D) When the light-tone stimulus had previously been associated with a water reward, BLA photo-

943 stimulation enhanced responding.

944

945 Figure 7. Bilateral infusions of d-amphetamine into the basolateral amygdala (BLA) intensify

946 the incentive value of a conditioned stimulus (CS). (A) In Exp. 4, rats received Pavlovian

947 conditioning. Bilateral cannulae were then implanted specifically into the BLA (‘BLA’ group) or into the

948 amygdala without targeting the BLA specifically (‘Amygdala’ group). (B) Estimated injector tip

949 placements in BLA rats and in Amygdala rats (anteroposterior position is shown in mm relative to

950 Bregma). (C, F) During Pavlovian conditioning, rats reliably learned the CS-unconditioned stimulus

951 contingency, as indicated by increasing CSR/PCSR ratios over sessions (ratio of nose-pokes into the

952 water receptacle made during each 5-second CS presentation versus during the 5-second period

953 preceding each CS presentation). Next, we assessed the effects of intra-cerebral d-amphetamine

954 infusions (0, 10 or 30 μg/hemisphere) on instrumental responding for the CS. Both (D, E) BLA and (G,

955 H) Amygdala rats pressed more on the active versus inactive lever, indicating that the CS acquired

956 incentive motivational value. (D) D-amphetamine influenced responding for the CS only when the drug

957 was infused into the BLA. n’s = 7-20/group. *p < 0.05.

A Pavlovian conditioning B Instrumental conditioning

CS Lights Tone Lights Tone Inactive UCS Water A Timeline ChR2-eYFP expression In vivo BLA/CeA B BLA C CeA surgery electrophysiology

≥ 4 weeks

D BLA E CeA H ChR2-expressing neurons 100 BLA 0.2 mV 75 50 ms

50 CeA 0.2 mV 50 ms 25

ChR2 fidelity (%) 20 Hz 0 110 20 40 110 20 40 BLA CeA F BLA G CeA 30 0.2 mV 0.2 mV 0.5 ms 0.5 ms

20 I eYFP-expressing neurons BLA 0.2 mV 10 50 ms CeA 0.2 mV Neuron firing (Hz) 0 50 ms 110 20 40 11020 40 Laser frequency (Hz) 20 Hz A Timeline BLA or CeA surgery Lights Tone Operant responding for photo-stimulation of the BLA or CeA

Left FR1 RR2 RR4 PR5 Extinction 5 vs 10 vs 20 Hz ≥ 4 weeks Right Reversal learning Uni = active lever = inactive lever B Histology C Stimulations D Active lever E Inactive lever 20 100 100 CeA Control BLA-ChR2 15 75 75 CeA-ChR2

10 50 50 - 1.8

5 * 25 * 25 - 2.12 Stimulations (30 min) 0 Lever presses (30 min) 0 Lever presses (30 min) 0 FR1 RR2 RR4 FR1 RR2 RR4 FR1 RR2 RR4 - 2.3 F PR5 G Extinction H frequency Active 80 75 # α 4 BLA * 60 α # OFF OFF 60 3 45 α 40 30 2 # - 1.8 15 *

Active lever #

20 Lever presses 1

presses (30 min) 0 15 Inactive *

- 2.12 Stimulations (1 min) 0 0 0 010203040 51020 - 2.3 CTL BLA Time (min) Frequency (Hz) CeA

I Left lever J Right lever K Unilateral - 2.56 150 α # 150 40 α # OFF OFF 30 * - 2.8 100 100

α # 20 - 3.14 50 50 10

- 3.3 Stimulations (30 min) Lever presses (30 min) 0 Lever presses (30 min) 0 0 -1 12 -1 12

Reversal Reversal CTL BLA CeA A Timeline BLA Pavlovian conditioning + surgery photo-stimulation

Lights Tone No laser CS CS Water ≥ 4 Paired weeks laser CS CS Unpaired laser CS CS

B Histology C Pavlovian conditioning

20 α # α # - 2.12 10 ChR2 6 # Paired - 2.3 laser Unpaired - 2.56 4 * laser

Control

- 2.8 CSR/PCSR ratio 2

- 3.14 0 - 3.3 1234 Sessions (2-day blocks) A Timeline BLA Pavlovian Instrumental conditioning + surgery conditioning photo-stimulation

Lights Tone Lights Tone No laser CS Water 5 Hz 10 Hz 20 Hz ≥ 4 Paired weeks laser CS CS CS

Unpaired 20 Hz laser CS

B No laser C Paired laser D Unpaired laser 90 90 90

75 75 75 ChR2 Active 60 60 60 Inactive

45 45 * 45 eYFP Active Inactive 30 30 30 α

Lever presses (20 min) 15 * 15 * 15 0 0 0 0 5 10 20 20 Laser frequency (Hz) Rat A Controls B ChR2 20 Hz 8 Rat Lights Tone 10 2 14 5 15 12 26 21 27 23

01015205 0 5 10 15 20 Time (minutes) Lights Tone Water C Controls Rat D ChR2 Rat 4 5 14 6 19 20 Hz 8 26 24 27 Lights Tone 25 29 38 33 41 34

0 5 10 15 20 0 5 10 15 20 Time (minutes) A Timeline B Histology Pavlovian BLA or amygdala D-amph microinjections BLA conditioning surgery + Instrumental conditioning Lights Tone Lights Tone d-amph Water - 1.8

- 2.12

BLA - 2.3 C Pavlovian conditioning Instrumental conditioning D Active E Inactive 6 300 - 2.56 5 * 4 200 - 2.8 3 - 3.14 2 100

CSR/PCSR ratio CSR/PCSR 1 - 3.3 0 Lever presses (40 min) 0 12345 0 10 30 0 10 30 Sessions d-amph dose (2-day blocks) (μg/hemisphere) Amygdala

Amygdala F Pavlovian conditioning Instrumental conditioning - 1.8 12 300 G Active H Inactive

9 - 2.12 200

6 - 2.3 100 3 - 2.56 CSR/PCSR ratio CSR/PCSR

0 Lever presses (40 min) 0 12345 0 10 30 0 10 30 - 2.8 Sessions d-amph dose (2-day blocks) (μg/hemisphere)