Cocaine produces widely recognized rewarding effects but also produces aversive effects that occur several minutes after rewarding effects dissipate. Prior work shows these aversive effects are particularly strong in some animals, in whom they produce conditioned avoidance effects that reduce overall cocaine-seeking. The sources of these individual variations are largely unknown, but we found evidence for contributions from heritable influences. Using outbred male and female Sprague Dawley and heterogeneous stock rats, we found that offspring of individuals expressing higher versus lower levels of conditioned avoidance to cocaine on a runway operant task are themselves higher or lower in this trait. These results did not differ between sexes and are consistent with a heritable influence driving conditioned avoidance of cocaine, which could either be genetic or nongenetic. Runway latency to obtain cocaine also differed markedly between seven inbred rat strains (tested in males only), again consistent with a heritable influence. Several control tasks showed that variations in cocaine avoidance were not explained by differences in exploratory locomotion, overall motivation, or resistance of food-seeking to footshock punishment. Notably, the latter punishment resistance task has been linked to addiction-like behaviors in its own right, and performance on this task also varied considerably between inbred strains but did so independently of cocaine avoidance. Hence, punishment resistance and cocaine avoidance may be influenced by independent heritable factors.

Addiction is a complex and heritable trait which progresses through several developmental stages, each of which is presumably influenced by multiple partially overlapping genetic factors. Cocaine initially produces rewarding effects, followed by aversive effects including anxiety, craving, anhedonia, and withdrawal. These aversive effects have been suggested to contribute to the etiology of cocaine use disorders (CUD), as repeated exposure is thought to desensitize the rewarding effects and sensitize the aversive effects through a process involving both aberrant reward-based learning and aberrant avoidance-based learning. We examined the genetic basis of aversion learning using both food-based and cocaine-based behavioral assays in outbred Heterogenous Stock (HS) rats. A total of 1,074 HS rats (35.3% male) underwent runway operant cocaine-seeking, food-based progressive ratio and punishment testing, and locomotion testing. These phenotypes were significantly heritable (with h2 estimates as high as 0.307) and identified significant (p < 0.05) genetic loci related to avoidance-based learning including from the punishment task on Chromosomes 2, 3, 5, and 6, and the cocaine-operant runway latency task on Chromosome X. 172 positional candidate genes were identified from significant and suggestive loci, including Cdh10, Cdh12, Cdh18 which have previously been associated with smoking initiation from human GWAS, Adcy3, Cfap206, and Drc1 which are associated with primary neuronal cilia, as well as SNPs associated with novelty-related and social interaction phenotypes in independent samples of HS rats. Our results suggest that these aversion learning phenotypes are themselves complex heritable traits influenced by multiple genetic loci, which may pleiotropically affect other aspects of addiction biology.

Genome-wide association studies typically evaluate the autosomes and sometimes the X Chromosome, but seldom consider the Y or mitochondrial (MT) Chromosomes. We genotyped the Y and MT Chromosomes in heterogeneous stock (HS) rats (Rattus norvegicus), an outbred population created from 8 inbred strains. We identified 8 distinct Y and 4 distinct MT Chromosomes among the 8 founders. However, only 2 types of each nonrecombinant chromosome were observed in our modern HS rat population (generations 81-97). Despite the relatively large sample size, there were virtually no significant associations for behavioral, physiological, metabolome, or microbiome traits after correcting for multiple comparisons. However, both Y and MT Chromosomes were strongly associated with the expression of a few genes located on those chromosomes, which provided a positive control. Our results suggest that within modern HS rats there are no Y and MT Chromosomes differences that strongly influence behavioral or physiological traits. These results do not address other ancestral Y and MT Chromosomes that do not appear in modern HS rats, nor do they address effects that may exist in other rat populations, or in other species.

Cocaine blocks dopamine reuptake, thereby producing rewarding effects that are widely studied. However, cocaine also blocks serotonin uptake, which we show drives, in rats, individually variable aversive effects that depend on serotonin 2C receptors (5-HT2CRs) in the rostromedial tegmental nucleus (RMTg), a major GABAergic afferent to midbrain dopamine neurons. 5-HT2CRs produce depolarizing effects in RMTg neurons that are particularly strong in some rats, leading to aversive effects that reduce acquisition of and relapse to cocaine seeking. In contrast, 5-HT2CR signaling is largely lost after cocaine exposure in other rats, leading to reduced aversive effects and increased cocaine seeking. These results suggest a serotonergic biological marker of cocaine-seeking vulnerability that can be targeted to modulate drug seeking.

Zhu-Tokita-Takenouchi-Kim (ZTTK) syndrome is a newly described autosomal dominant multisystem developmental disorder resulting from a mutation of the SON gene located on chromosome region 21q22.11. It is characterized by heterogeneous features such as intellectual disability, facial dysmorphisms, poor feeding, vision abnormalities, musculoskeletal anomalies, congenital heart and genitourinary system defects, as well as several unique neurological findings including seizures, tone abnormalities, autism spectrum disorder and variable brain abnormalities noted on neuroimaging. Unfortunately, we lack adequate information regarding the spectrum of these neurological symptoms. In this study, we report 2 new unrelated cases of ZTTK syndrome, and identify new pathogenic variants in the SON gene through microarray analysis and whole-exome sequencing. We also emphasize the neurological manifestations of the syndrome in our patients and discuss the significance of gathering more data regarding neurological presentation, particularly seizure characteristics and long-term developmental progression. This information will be crucial to help understand long-term neurodevelopmental prognosis in these patients.

Many studies have examined the mechanisms by which drugs of abuse produce

Alawieh, Ali M MD, PhD; Eid, Maya BS; Anadani, Mohammad MD; Sattur, Mithun MD; Maier, Ilko L MD; Feng, Wuwei MD, MS; Goyal, Nitin MD; Starke, Robert M MD; Rai, Ansaar MD; Fargen, Kyle M MD, MPH; Psychogios, Marios-Nikos MD; De Leacy, Reade MD; Grossberg, Jonathan A MD; Keyrouz, Salah G MD; Dumont, Travis M MD; Kan, Peter MD; Lena, Jonathan MD; Liman, Jan MD; Arthur, Adam S MD, MPH; Elijovich, Lucas MD; Mccarthy, David J MD; Saini, Vasu MD; Wolfe, Stacey Q MD; Mocco, J MD; Fifi, Johanna T MD; Nascimento, Fábio A MD; Giles, James A MD; Allen, Michelle , BSN; Crosa, Roberto MD; Fox, W Christopher MD; Gory, Benjamin MD; Spiotta, Alejandro M MD, on behalf of Stroke Thrombectomy and Aneurysm Registry (STAR) Collaborators

Although cocaine is powerfully rewarding, not all individuals are equally prone to abusing this drug. We postulate that these differences arise in part because some individuals exhibit stronger aversive responses to cocaine that protect them from cocaine seeking. Indeed, using conditioned place preference (CPP) and a runway operant cocaine self-administration task, we demonstrate that avoidance responses to cocaine vary greatly between individual high cocaine-avoider and low cocaine-avoider rats. These behavioral differences correlated with cocaine-induced activation of the rostromedial tegmental nucleus (RMTg), measured using both in vivo firing and c- fos , whereas slice electrophysiological recordings from ventral tegmental area (VTA)-projecting RMTg neurons showed that relative to low avoiders, high avoiders exhibited greater intrinsic excitability, greater transmission via calcium-permeable AMPA receptors (CP-AMPARs), and higher presynaptic glutamate release. In behaving animals, blocking CP-AMPARs in the RMTg with NASPM reduced cocaine avoidance. Hence, cocaine addiction vulnerability may be linked to multiple coordinated synaptic differences in VTA-projecting RMTg neurons. SIGNIFICANCE STATEMENT Although cocaine is highly addictive, not all individuals exposed to cocaine progress to chronic use for reasons that remain unclear. We find that cocaine's aversive effects, although less widely studied than its rewarding effects, show more individual variability, are predictive of subsequent propensity to seek cocaine, and are driven by variations in RMTg in response to cocaine that arise from distinct alterations in intrinsic excitability and glutamate transmission onto VTA-projecting RMTg neurons.

BACKGROUND Randomized controlled trials evaluating mechanical thrombectomy (MT) for acute ischemic stroke predominantly studied anterior circulation patients. Both procedural and clinical predictors of outcome in posterior circulation patients have not been evaluated in large cohort studies. OBJECTIVE To investigate technical and clinical predictors of functional independence after posterior circulation MT while comparing different frontline thrombectomy techniques. METHODS In a retrospective multicenter international study of 3045 patients undergoing MT for stroke between 06/2014 and 12/2018, 345 patients had posterior circulation strokes. MT was performed using aspiration, stent retriever, or combined approach. Functional outcomes were assessed using the 90-d modified Rankin score dichotomized into good (0-2) and poor outcomes (3-6). RESULTS We included 2700 patients with anterior circulation and 345 patients with posterior circulation strokes. Posterior patients (age: 60 ± 14, 46% females) presented with mainly basilar occlusion (80%) and were treated using contact aspiration or ADAPT (39%), stent retriever (31%) or combined approach (19%). Compared to anterior strokes, posterior strokes had delayed treatment (500 vs 340 min, P < .001), higher national institute of health stroke scale (NIHSS) (17.1 vs 15.7, P < .01) and lower rates of good outcomes (31% vs 43%, P < .01). In posterior MT, diabetes (OR = 0.28, 95%CI: 0.12-0.65), admission NIHSS (OR = 0.9, 95%CI: 0.86-0.94), and use of stent retriever (OR = 0.26, 95%CI: 0.11-0.62) or combined approach (OR = 0.35, 95%CI: 0.12-1.01) vs ADAPT were associated with lower odds of good outcome. Stent retriever use was associated with lower odds of good outcomes compared to ADAPT even when including patients with only basilar occlusion or with successful recanalization only. CONCLUSION Despite similar safety profiles, use of ADAPT is associated with higher rates of functional independence after posterior circulation thrombectomy compared to stent retriever or combined approach in large "real-world" retrospective study.

Introduction: Aspiration thrombectomy using the ADAPT technique has been shown to have similar efficacy to stent retriever thrombectomy (SRT) in randomized trials of proximal large vessel occlusions. In this work, we investigated the differences in technical and clinical outcomes between ADAPT and SRT for distal vessel occlusions from the Stroke Thrombectomy and Aneurysm Registry (STAR). Methods: Patients undergoing thrombectomy for acute ischemic stroke at 12 comprehensive stroke centers in the US and Europe between 01/2013 and 12/2018 were reviewed. Data was collected retrospectively from patient charts, procedure notes, and patient follow-up in neurology clinics for patients with isolated distal artery occlusion including MCA2, MCA3/4, ACA1/2, and PCA2/3. Clinical endpoint was the modified Rankin score (mRS) at 90-days, and technical outcomes were procedure time, total attempts, and mTICI scores. Results: A total of 464 patients (mean age 69±13.5 years) were treated with ADAPT (56%) or SRT (44%) for distal occlusions during the study period. Patients in the ADPAT group were mainly treated using 3MAX (36%), 4MAX (21%), ACE68/64 (20%), 5MAX/ACE (12%). SRT group included the use of Trevo (50%), Solitaire (44%), or both (5%). There were no significant differences in rates of good outcomes or successful recanalization between ADAPT and SRT groups on multivariate logistic regression analysis controlling for significant confounding variables (p>0.1). Use of SRT in distal occlusions was an independent predictor of longer procedure times compared to ADAPT on linear regression (coefficient=23, p<0.001), and there was a trend toward higher odds of symptomatic hemorrhage in the SRT group (OR=2.6, p=0.06) on multivariate analysis. There were no differences in mortality and complication rates between the two groups. Conclusions: Both SRT and ADAPT thrombectomy lead to comparable rates of favorable outcome for distal vessel occlusion. SRT requires longer procedures and may be associated with higher rates of hemorrhage. Further randomized trials are needed to confirm whether either techniques may provide a better safety or efficacy profile in distal vessel occlusions.

The aversive properties associated with drugs of abuse influence both the development of addiction and relapse. Cocaine produces strong aversive effects after rewarding effects wear off, accompanied by increased firing in the lateral habenula (LHb) that contributes to downstream activation of the rostromedial tegmental nucleus (RMTg). However, the sources of this LHb activation are unknown, as the LHb receives many excitatory inputs whose contributions to cocaine aversion remain uncharacterized. Using cFos activation and in vivo electrophysiology in male rats, we demonstrated that the rostral entopeduncular nucleus (rEPN) was the most responsive region to cocaine among LHb afferents examined and that single cocaine infusions induced biphasic responses in rEPN neurons, with inhibition during cocaine's initial rewarding phase transitioning to excitation during cocaine's delayed aversive phase. Furthermore, rEPN lesions reduced cocaine-induced cFos activation by 2-fold in the LHb and by a smaller proportion in the RMTg, while inactivation of the rEPN or the rEPN-LHb pathway attenuated cocaine avoidance behaviors measured by an operant runway task and by conditioned place aversion (CPA). These data show an essential but not exclusive role of rEPN and its projections to the LHb in processing the aversive effects of cocaine, which could serve as a novel target for addiction vulnerability. SIGNIFICANCE STATEMENT Cocaine produces well-known rewarding effects but also strong aversive effects that influence addiction propensity, but whose mechanisms are poorly understood. We had previously reported that the lateral habenula (LHb) is activated by cocaine and contributes to cocaine's aversive effects, and the current findings show that the rostral entopeduncular nucleus (rEPN) is a major contributor to this LHb activation and to conditioned avoidance of cocaine. These findings show a critical, though not exclusive, rEPN role in cocaine's aversive effects, and shed light on the development of addiction.

Introduction: Recently completed randomized controlled trials comparing aspiration thrombectomy (ADAPT) to stent retriever thrombectomy (SRT) demonstrated similar clinical outcomes, but faster thrombectomy procedure time in the ADAPT group. This study evaluates the difference in technical outcomes between ADAPT and SRT combined with balloon-guide catheters (BGC). Methods: Patients undergoing thrombectomy for acute ischemic stroke at 12 comprehensive stroke centers in the US and Europe between 01/2013 and 12/2018 were reviewed. Data was collected retrospectively from patient charts, procedure notes, and patient follow-up in neurology clinics. Clinical endpoint was the modified Rankin score (mRS) at 90-days, and technical outcomes were procedure time, total attempts, and mTICI scores. Results: The study included 2,016 patients (mean age 69±15) who underwent stroke thrombectomy using ADAPT (46%), SRT (46%), or SRT+BGC (8%). Similar baseline characteristics were observed between the three groups, and no significant difference in mRS scores at 90-days between the three groups in univariate and multivariate analyses. Thrombectomy performed using SRT+BGC required significantly shorter procedure time compared to SRT (35 vs 61 min, p<0.001) that was comparable to ADAPT (36 min, p>0.1). However, use of SRT+BGC required significantly lower number of aspiration attempts compared to ADAPT (median 1 vs. 2, p<0.05). On multivariate linear regression, use of SRT+BGC independent predicted a significant reduction in procedure time compared to SRT (coefficient=-30.6, p<0.001), and significantly lower number of attempts compared to ADAPT (coefficient=-0.4, p=0.01). SRT+BGC was an independent predictor of higher mortality compared to ADAPT (OR=2.4, p<0.01), despite comparable rates of favorable outcomes (mRS 0-2) between the two groups. Use of SRT+BGC was not an independent predictor of symptomatic hemorrhage or complications compared to SRT or ADAPT. Conclusions: This study shows that although ADAPT allows for faster procedure time compared to SRT, the use of BGC in SRT allows for a comparable procedure time to ADAPT with similar overall rates of favorable outcome, complications and hemorrhage. Mortality was higher with the use of BGC compared to ADAPT.

Introduction: Elderly patients, octogenarians and nonagenarians, were excluded or under-represented in the majority of stroke endovascular thrombectomy (ET) trials. There is conflicting data on the outcomes of ET in the elderly. We evaluated age-dependent outcomes of ET for stroke in a large dataset from the Stroke Thrombectomy and Aneurysm Registry (STAR). Methods: Patients undergoing ET for acute ischemic stroke at 12 comprehensive stroke centers in the US and Europe between 01/2013 and 12/2018 were reviewed. Data was collected retrospectively from patient charts, procedure notes, and patient follow-up in neurology clinics. The primary endpoint was the modified Rankin score (mRS) at 90-days which was dichotomized into good outcome (mRS 0-2) or poor outcome (mRS 3-6). Results: Of 3,850 patients reviewed, 2,827 had 90-day follow-up (mean age 69±14), and were divided into 6 age groups: 20-49 (G1, 10%), 50-59 (G2, 10%), 60-69 (G3, 23%), 70-79 (G4, 27%), 80-89 (G5, 21%), 90 or more (G6, 4%). When adjusted for confounding variables, age was an independent predictor of poor outcome (OR=1.4, p<0.001) and mortality (OR=1.5, p<0.0001). When used as categorical variable, adjusted OR (aOR) for good outcomes were significantly lower in groups G2-G6 compared to G1 (p<0.01, figure), and OR for mortality were significantly higher in G2-G6 compared to G1 (p<0.01, figure). An age increment of 10 years was associated with 23% higher odds of symptomatic hemorrhage, and 50% higher odds of mRS 5-6. The impact of procedure time on good outcome (mRS 0-2) was also age-dependent with aOR=0.84 (p<0.05) in G1,2 compared to aOR=0.65 (p<0.05) in G5,6. Conclusions: Age is a major predictor of functional recovery after ET, and this study demonstrates a clear age-dependent increase in rate of mortality and poor outcomes after ET with exponentially worse outcomes above 80 years of age. Complication rates were not age-dependent. Further studies are required to optimize patient selection for ET in the elderly.

Introduction: Endovascular thrombectomy (ET) for treating acute stroke in the real-world has expanded beyond the selection criteria used in major trials, and currently includes posterior circulation strokes. Posterior circulation stroke is believed to have worse outcomes than anterior circulation stroke, and its outcomes following ET are still being studied. We explored the major determinants of functional recovery after ET for posterior circulation stroke in a large cohort of patients from the Stroke Thrombectomy and Aneurysm Registry (STAR). Methods: STAR includes patients undergoing ET for acute ischemic stroke at 12 comprehensive stroke centers in the US and globally. Data on patient demographics, technical and clinical outcomes was reviewed retrospectively from patient charts and procedure notes. Primary outcomes was the modified Ranking Score (mRS) at 90 days dichotomized into favorable (mRS 0-2) and poor outcome (mRS 3-6). Results: A total of 3850 patients were reviewed, of which 345 patients (mean age 60±14) were treated for posterior circulation stroke with predominantly basilar artery occlusion (80%). Patients were treated using aspiration thrombectomy (ADAPT, 39%), stent retriever thrombectomy (31%), combined approach (19%) or intracranial stenting (7%). The overall rate of favorable outcome was 33%. Patients with diabetes, high NIHSS on admission, and proximal occlusions had significantly higher odds of poor functional outcomes on multivariate analysis (p<0.05). Compared to ADAPT thrombectomy, significantly higher odds for poor outcomes were observed with the use of stent retriever (aOR=0.84, p<0.01) or primary combined approach (aOR=2.85, p=0.05). The advantage of ADAPT on functional recovery compared to stent retrievers persisted when regression models were limited to patients with successful recanalization, or with basilar artery occlusions. No differences in complication and hemorrhage rates were observed. Conclusions: Despite similar rates of functional recovery after ET for anterior circulation stroke between stent retriever and ADAPT, our analysis demonstrates that in posterior circulation stroke, ADAPT may lead to better functional outcomes compared to stent retriever without differences in safety profiles.

Persistence of reward seeking despite punishment or other negative consequences is a defining feature of mania and addiction, and numerous brain regions have been implicated in such punishment learning, but in disparate ways that are difficult to reconcile. We now show that the ability of an aversive punisher to inhibit reward seeking depends on coordinated activity of three distinct afferents to the rostromedial tegmental nucleus (RMTg) arising from cortex, brainstem, and habenula that drive triply dissociable RMTg responses to aversive cues, outcomes, and prediction errors, respectively. These three pathways drive correspondingly dissociable aspects of punishment learning. The RMTg in turn drives negative, but not positive, valence encoding patterns in the ventral tegmental area (VTA). Hence, punishment learning involves pathways and functions that are highly distinct, yet tightly coordinated.

Background Although cocaine is powerfully rewarding, only a fraction of drug-exposed individuals actually transition to become drug-dependent users, possibly due in part to genetic influences, as addictions are among the most heritable of human neuropsychiatric disorders. Consistent with addiction being a complex disorder, prior studies have noted many factors that predict addiction vulnerability, such as drug-induced aversive effects, behavioral responses to novelty, and sensitivity to punishment.Methods We tested different inbred strains of rats, as well as selectively bred Sprague-Dawley (SD) or Heterogeneous Stock (HS) rats, on cocaine avoidance, sensitivity to punishment, and locomotor responses to novelty, and calculated heritability estimates of these behaviors. We tested animals on additional control tasks (progressive ratio, shock avoidance) to control for alternate interpretations of addiction-like behaviors.Results Conditioned avoidance to cocaine varied greatly between individual rats as measured using either the runway operant task or conditioned place aversion (CPA). Cocaine avoidance responded rapidly to selective breeding, yielding heritability estimates of 0.70 and 0.58 in SD and HS rats. Resistance to punishment was also highly heritable in inbred rats (estimated h2 = 0.62), and varied independently of cocaine avoidance despite both behaviors being mediated by the RMTg. Furthermore, cocaine avoidance correlated positively with cocaine-induced c-Fos induction in the RMTg, negatively with initial rates of acquisition of intravenous cocaine self-administration, and positively with both cued and cocaine-primed reinstatement.Conclusions Cocaine avoidance and resistance to punishment are strongly and independently heritable behaviors that may both control different aspects of individual propensity to acquire and maintain drug-seeking.

The rostromedial tegmental nucleus (RMTg), a GABAergic afferent to midbrain dopamine (DA) neurons, has been hypothesized to be broadly activated by aversive stimuli. However, this encoding pattern has only been demonstrated for a limited number of stimuli, and the RMTg influence on ventral tegmental (VTA) responses to aversive stimuli is untested. Here, we found that RMTg neurons are broadly excited by aversive stimuli of different sensory modalities and inhibited by reward-related stimuli. These stimuli include visual, auditory, somatosensory and chemical aversive stimuli, as well as "opponent" motivational states induced by removal of sustained rewarding or aversive stimuli. These patterns are consistent with broad encoding of negative valence in a subset of RMTg neurons. We further found that valence-encoding RMTg neurons preferentially project to the DA-rich VTA versus other targets, and excitotoxic RMTg lesions greatly reduce aversive stimulus-induced inhibitions in VTA neurons, particularly putative DA neurons, while also impairing conditioned place aversion to multiple aversive stimuli. Together, our findings indicate a broad RMTg role in encoding aversion and driving VTA responses and behavior.

Article Figures and data Abstract Introduction Results Discussion Materials and methods Data availability References Decision letter Author response Article and author information Metrics Abstract The rostromedial tegmental nucleus (RMTg), a GABAergic afferent to midbrain dopamine (DA) neurons, has been hypothesized to be broadly activated by aversive stimuli. However, this encoding pattern has only been demonstrated for a limited number of stimuli, and the RMTg influence on ventral tegmental (VTA) responses to aversive stimuli is untested. Here, we found that RMTg neurons are broadly excited by aversive stimuli of different sensory modalities and inhibited by reward-related stimuli. These stimuli include visual, auditory, somatosensory and chemical aversive stimuli, as well as “opponent” motivational states induced by removal of sustained rewarding or aversive stimuli. These patterns are consistent with broad encoding of negative valence in a subset of RMTg neurons. We further found that valence-encoding RMTg neurons preferentially project to the DA-rich VTA versus other targets, and excitotoxic RMTg lesions greatly reduce aversive stimulus-induced inhibitions in VTA neurons, particularly putative DA neurons, while also impairing conditioned place aversion to multiple aversive stimuli. Together, our findings indicate a broad RMTg role in encoding aversion and driving VTA responses and behavior. https://doi.org/10.7554/eLife.41542.001 Introduction The rostromedial tegmental nucleus (RMTg), also called the tail of the ventral tegmental area (tVTA), is a GABAergic nucleus first described in 2009 as a major inhibitory input to neurons in the ventral tegmental area (VTA) (Jhou et al., 2009; Kaufling et al., 2009). Previous studies have shown that RMTg axons overwhelmingly synapse on tyrosine hydroxylase (TH) positive neurons in the VTA and substantia nigra, and electrical stimulation of the RMTg dramatically suppresses DA neuron firing (Balcita-Pedicino et al., 2011; Bourdy et al., 2014). Subsequent studies have further suggested an important RMTg role in conveying information about aversive stimuli onto VTA and substantia nigra DA neurons and mediating behavioral responses to these aversive stimuli (Brown et al., 2017; Hong et al., 2011; Jhou et al., 2009; Jhou et al., 2013; Vento et al., 2017). However, some fundamental questions about the proposed RMTg role in aversive encoding have not been addressed. Most prior studies had only tested RMTg firing responses to a limited range of aversive stimuli, for example footshocks, airpuffs, and their predictive cues (Hong et al., 2011; Jhou et al., 2009). One notable exception is a recent study that examined 11 distinct aversive stimuli, but found that only 3 of them increased RMTg expression of the immediate-early gene c-Fos (Sánchez-Catalán et al., 2017). Because c-Fos is often used as a proxy for neuronal firing, this result raises questions about the generalizability of RMTg responses to other aversive stimuli, albeit with the caveat that c-Fos likely reflects intracellular signaling events rather than firing per se (Kovács, 2008). In addition to their limited range of tested stimuli, earlier studies had also failed to characterize the influence of RMTg neurons on DA responses to aversive stimuli, raising further questions about the proposed RMTg role. Increasingly many studies have noted that DA responses to aversive stimuli, although somewhat complex, often exhibit a prominent inhibitory component (Fiorillo et al., 2013; Matsumoto et al., 2016; Tian and Uchida, 2015), and that the aversive stimulus-induced inhibition is associated with behavioral avoidance and diminished learning (Chang et al., 2016; Chang et al., 2018; Lammel et al., 2012). Previous study has suggested that these inhibitory responses are primarily driven by GABAergic transmissions onto DA neurons (Henny et al., 2012). However, the sources of the GABAergic transmissions are still unknown, and have been proposed to arise from numerous possible sources including not only the RMTg, but also the lateral hypothalamus, ventral pallidum, extended amygdala, and GABAergic neurons in the VTA (Jennings et al., 2013; Tian et al., 2016). Hence, the specific RMTg contribution to VTA responses is not known. To address these questions about generalizability of RMTg responses to aversive stimuli and its influence on DA neurons, we recorded VTA and RMTg neuron responses to a wide range of aversive stimuli in freely-moving rats, while using excitotoxic lesions to investigate the RMTg influence on VTA neuron firings and on conditioned place aversion to these stimuli. Results RMTg neurons are activated by diverse phasic aversive stimuli To examine the generalizability of RMTg responses to aversive stimuli, we recorded RMTg neuron responses to six distinct aversive stimuli covering a broad range of stimulus modalities in animals that had been previously trained to distinguish auditory cues predicting reward delivery or nothing (Figure 1A,B). Three of these stimuli were phasic: footshock (0.7 mA), loud siren (115 dB), and bright light (1600 lumens), which lasted for 10 ms, 1 s, and 2 s respectively. Three additional stimuli were sustained: lithium chloride (LiCl), restraint stress, and cocaine, lasting for several minutes in duration (Figure 2A). Stimuli were chosen to represent distinct sensory modalities, whose aversive qualities had been noted previously (Ettenberg et al., 1999; Tzschentke, 2007; Winston et al., 2001). All recorded neurons were also tested for their response to reward cues and neutral tones (75 dB) that were followed by a sucrose pellet delivery and no consequence, respectively. Figure 1 Download asset Open asset Behavioral training procedure. (A) Schematic of training paradigm. (B) Training performance. Response index is the percentage of trials in which animals made nose pokes within 2 s of onset of reward or neutral cues. https://doi.org/10.7554/eLife.41542.002 Figure 2 with 1 supplement see all Download asset Open asset RMTg neurons are activated by diverse phasic aversive stimuli. (A) Schematic of recording paradigm. (B) Raster plots of a representative RMTg neuron response to reward cues and footshocks. (C) RMTg neurons on average showed inhibition to reward-predictive cues during a 200–400 ms post-stimulus window, and small excitations to neutral tones and large excitations to footshock, siren, and bright light during a 0–100 ms post-stimulus windows. (D) Percentage of RMTg neurons that showed inhibition, excitation or no response to stimuli (200–400 ms window for reward cues, and 0–100 ms window for aversive or neutral stimuli). (E) Scatterplot of individual neurons’ responses to reward cues and footshocks. Many reward-cue inhibited neurons were also excited by footshocks, consistent with a valence-encoding pattern. Blue solid dots: neurons significantly responding to both reward cues and footshocks. Gray shaded box: neurons inhibited by reward cues and excited by footshocks, consistent with hypothesized valence-encoding. (F, G) RMTg neurons activated by footshocks tended to also be activated by siren and bright light, in proportion to the magnitude of response to footshock. Solid dots: neurons significantly responding to siren and bright light. * indicates p < 0.05, ** p < 0.01, *** < 0.0001. https://doi.org/10.7554/eLife.41542.003 Out of 151 recorded neurons, 59 were located in the RMTg, as defined by immunostaining for FOXP1 (Lahti et al., 2016; Smith et al., 2018) (Figure 2—figure supplement 1A,B). Consistent with previous studies that RMTg neurons encode motivational valence (Hong et al., 2011; Jhou et al., 2009), we found that RMTg neurons on average showed significant inhibition to reward cues during a time window 200–400 ms post-stimulus, and rapid excitations to all other phasic stimuli 0–100 ms post-stimulus (p < 0.0001 for all stimuli) (Figure 2B; Figure 2C). Moreover, responses to the three phasic aversive stimuli (footshock, siren, and bright light) were significantly greater than responses to the neutral tone, suggesting an aversion-related augmentation in RMTg responses to aversive stimuli (F = 1.582, p = 0.027, p = 0.0001, and p = 0.0048 for neutral tone compared with footshock, siren, and bright light, repeated measures one-way ANOVA, Holm-Sidak test for multiple comparison) (Figure 2C). Notably, even though the duration of the aversive stimuli ranged from 10 ms up to 2 s, RMTg neuron response durations were largely confined to a period 0–100 ms after stimulus onset, independently of the total stimulus duration. When individual neurons (instead of population averages) were analyzed, we observed considerable variation among individual neuron responses to these stimuli. Specifically, we found 51% (30/59) of RMTg neurons showing significant changes relative to baseline firing (in either direction) to reward cue during a window 200–400 ms post-stimulus, with almost three-fourths of responsive neurons (22/30 = 73%) showing inhibition, and the remaining showing excitation. Furthermore, 59% (35/59), 70% (41/59), and 61% (36/59) of RMTg neurons showed significant changes from baseline firing (in either direction) to footshock, siren, and bright light during a window 0–100 ms post-stimulus, again with roughly three-fourths showing excitations (24/35 = 68%, 31/41 = 75%, and 26/36 = 72% for footshock, siren, and light, respectively) (Figure 2D). Interestingly, we noticed that a majority (14/22) of reward cue-inhibited neurons showed significant excitations to footshock, even though shock-excited neurons are only 41% of all RMTg neurons, indicating that shock-excitation patterns are particularly concentrated among the reward-cue inhibited population, contributing to an overall valence-encoding pattern in this population (Figure 2E). Moreover, among RMTg neurons significantly excited by footshock, these responses were positively correlated with responses of each individual neuron to siren and bright light, indicating that neurons more strongly activated by the footshock were also more strongly activated by siren and bright light (r2 = 0.2252, p = 0.004 and r2 = 0.1547, p = 0.0194 for siren and light respectively, 0–100 ms post-stimulus) (Figure 2F,G). RMTg neurons exhibit biphasic responses to sustained aversive stimuli consistent with opponent process theory In addition to testing RMTg responses to multiple phasic aversive stimuli, we also examined these same neurons’ responses to one of several sustained aversive stimuli lasting several minutes each. These were: a low dose of LiCl (10 mg/kg i.p.), restraint stress (6 min), or cocaine (0.75 mg/kg i.v.), which previous studies had shown produce rewarding effects for about 10 min followed by an aversive ‘crash’ beginning around 15 min post-injection (Ettenberg et al., 1999; Jhou et al., 2013). All stimuli were administered roughly 20 min after rats had been tested with phasic stimuli (Figure 2A), allowing subsequent comparison of neural responses across multiple stimuli. Notably, the dose of LiCl that we used is relatively low compared to the much higher doses commonly used to induce conditioned taste aversions, and is thought to produce modest aversive effects lasting only approximately 15 min (Tomasiewicz et al., 2006). All these sustained stimuli produced two distinct phases of response in RMTg firing. We found that both LiCl (n = 22) and restraint stress (n = 12) increased RMTg firing for several minutes after the onset of each stimulus (first 0–10 min and 0–3 min post-stimulus, respectively, p = 0.01 and p = 0.02, repeated measure one-way ANOVA, Holm-Sidak test for multiple comparison). These periods of activation are somewhat shorter than the expected duration of the aversion for each stimulus (15 and 6 min, respectively), suggesting some habituation of the RMTg excitation even while the stimulus is still ongoing. Interestingly, RMTg firing exhibiting a ‘rebound’ inhibition to below baseline levels just after the offset of the aversive phases of each stimulus (20–30 min post-stimulus for LiCl and 9–12 min for restraint stress, p = 0.047, and p = 0.036, respectively, repeated measure one-way ANOVA, Holm-Sidak test for multiple comparison) (Figure 3A,E). In contrast, saline injections of equal volume as the LiCl injections had no effect on RMTg firing during either of the two time windows where LiCl had produced responses (n = 12, p = 0.23 and p = 0.84, respectively) (Figure 3A). As with phasic stimuli, analysis of individual neurons showed heterogeneous responses to LiCl and restraint stress. Roughly 36–42% of RMTg neurons showed both an initial excitation and a rebound inhibition (8/22 and 5/12 neurons for LiCl and restraint stress, respectively), while the remaining neurons typically showed responses during only one of the two phases (Figure 3B,F). We conducted further analyses to examine whether responses during initial and rebound phases were consistent with encoding of motivational states. We found that RMTg responses during the initial phase of LiCl and restraint stress were positively correlated with responses of these same neurons to footshock (r2 = 0.1469, p = 0.0033 and r2 = 0.3783, p = 0.033, respectively) (Figure 3C,G), while their rebound inhibitory phase responses correlated with responses to reward cues (r2 = 0.2165, p < 0.0001 and r2 = 0.3954, p = 0.038, respectively) (Figure 3D,H). Figure 3 Download asset Open asset RMTg neurons exhibit biphasic responses to sustained aversive stimuli consistent with opponent process theory. (A, B) RMTg neurons showed activation to a low dose of LiCl for 10 min post injection, and (E, F) to restraint stress during the first 3 min of the 6 min restraint. Both aversive stimuli also produced a rebound inhibition of firing below baseline during a later time window (20–30 min window for LiCl, and 9–12 min window for restraint stress). (C, D) Individual neuron responses during initial phases of LiCl and restraint stress correlated with their responses to footshock, while responses during rebound phases correlated with responses to reward cue (G, H). (I, J) Cocaine infusion produced an opposing pattern in RMTg neurons, with inhibition during the first 10 min followed by a rebound excitation 15–25 min post infusion. (K, L) Individual responses during rebound (aversive) phase of cocaine were correlated with their responses to footshock, while responses during initial (rewarding) phase were correlated with their responses to reward cue. (M) Schematic of conditioned place preference regimen. (N) Low dose of LiCl (10 mg/kg) i.p. injection induced place aversion during 0–15 min and place preference during 15–30 min post-stimulus. (O) Cocaine (0.75 mg/kg) i.v. infusion induced place preference during 0–15 min and place aversion 15–30 min post-stimulus. (P) Saline injection did not produce place preference during either time window. Solid dots in B, F, and J: neurons significantly responding during both initial and rebound phases. https://doi.org/10.7554/eLife.41542.005 As LiCl and restraint stress are both aversive, we hypothesized that the rebound inhibitions of RMTg neurons may be correlated with the rewarding ‘relief’ upon the removal of these stimuli. Conversely, we hypothesized that the converse might also be true, that RMTg neurons might show a rebound excitation after removal of a sustained rewarding stimulus, as we had observed earlier in the LHb (Jhou et al., 2013). Thus, we further examined RMTg responses to a 0.75 mg/kg i.v. cocaine infusion. We found that RMTg neurons (n = 38) showed similar bi-phasic responses to cocaine as to LiCl and restraint stress, but in the opposite direction. Specifically, they were initially inhibited by cocaine 0–10 min post-infusion (initial phase), and subsequently activated 20–30 min post-infusion (rebound phase), when cocaine was aversive (p = 0.0001 and p = 0.038, repeated measure one-way ANOVA, Holm-Sidak test for multiple comparison). Saline infusions (n = 14) had no effect on RMTg firing during either phase (p = 0.62 and p = 0.25) (Figure 3I). Furthermore, when individual neuron responses were analyzed, we found that 32% of RMTg neurons (12/38) exhibited both inhibition during the initial phase and excitation during the rebound phase, again consistent with the negative valence-encoding pattern seen previously (Figure 3J). Furthermore, responses during the initial phase were positively correlated with responses of these same neurons to food-predictive cues (r2 = 0.3267, p < 0.0001), while responses during the rebound phase were positively correlated with responses to footshocks (r2 = 0.1469, p = 0.003). Although these results imply that RMTg neurons bi-directionally encode aversive and rewarding properties of sustained stimuli, they do not indicate whether the early and rebound phases of LiCl and cocaine are indeed aversive and rewarding. Thus, we tested four groups of animals (eight per group) for conditioned place preference or aversion to the same doses of LiCl and cocaine as during the recordings. We placed animals into conditioning chambers either 0–15 min or 15–30 min post-stimulus, to align with the two phases of neural responses seen in earlier recordings (Figure 3M). In separate groups of LiCl-treated groups, we found that animals developed place aversion when placed into conditioning chambers immediately after the injection, but showed place preference when placed into chambers 15 min after the injection (p < 0.0001 and p = 0.02 for 0–15 min and 15–30 min, respectively) (Figure 3N). In contrast, separate groups of cocaine-treated animals developed place preference if conditioned immediately after infusion, and place aversion if conditioned 15 min later (p = 0.034 and p < 0.0001, respectively) (Figure 3O). Saline injections did not produce place preference during either time window (p = 0.412 and p = 0.562, respectively) (Figure 3P). Together, our results indicate that bidirectional behavioral effects of sustained rewarding or aversive stimuli correlate with bidirectional responses of RMTg neurons to these stimuli, suggesting a possible substrate for an ‘opponent’ responses to stimuli described decades ago by opponent process theory (Solomon and Corbit, 1973). VTA-projecting RMTg neurons preferentially show valence-encoding patterns As previous results have indicated heterogeneities in RMTg responses to both phasic and sustained stimuli, with only 30–40% of RMTg neurons encoding negative valence, we next examined whether the heterogeneity in response patterns might be related to heterogeneity in RMTg projection targets. Thus, we used endoscopic calcium imaging and selectively recorded RMTg neurons that project to either the VTA or the dorsal raphe nucleus (DRN), both of which are implicated in encoding of motivational stimuli (Cohen et al., 2012; Li et al., 2016). Because of the need to keep GRIN lenses short, this experiment was performed in mice, rather than rats, but the anatomy of the RMTg is very similar between species (Smith et al., 2018). We injected into wild type mice a retrogradely transported canine adenovirus expressing Cre recombinase (CAV2-Cre) into either the VTA or (in separate mice) the DRN, along with a second virus into the RMTg expressing a Cre-dependent fluorescent calcium indicator (gCaMP6f) (Figure 4A–C, Figure 4—figure supplement 1A,B). Mice were tested in separate sessions in which they received either a footshock (0.3mA), or an auditory tone followed by food pellet delivery. Figure 4 with 1 supplement see all Download asset Open asset VTA-projecting RMTg neurons preferentially show valence-encoding patterns. (A) Cav2-cre injected into the VTA or DRN was retrogradely transported to the RMTg in mice, driving gCaMP6f expression in subsets of RMTg neurons projecting to the VTA or DRN respectively. (B) Representative photograph of the RMTg region in which gCaMP6f in RMTg (green label) is co-expressed with FOXP1 (red), a transcription factor locally specific to RMTg neurons. (C) A representative photo of gCaMP6f positive neurons in vivo (upper panel) and denoised Ca2+ traces extracted from the marked neurons (lower panel). (D) VTA-projecting RMTg neurons showed an average inhibition by reward cues, and excitation by footshocks. (E) DRN-projecting neurons showed no average response to reward cues, but were excited by footshocks. (F) Among VTA-projecting RMTg neurons, neurons showing stronger excitations to shock tended to also show stronger inhibitions to the reward cue, while individual DRN neurons did not show this correlation (G). Colored solid dots: neurons significantly responding to both reward cues and footshocks. (H) VTA-projecting neurons were much more likely to be inhibited by the reward cue than DRN-projecting neurons, and much less likely to be activated, while VTA- and DRN-projecting neurons were both predominantly activated by shock. https://doi.org/10.7554/eLife.41542.006 We found that VTA-projecting neurons were on average inhibited by reward cues and activated by footshocks, that is a negative valence-encoding pattern (Figure 4D, Figure 4—figure supplement 1C,D). Analysis of individual neurons further confirmed this; among 25 VTA-projecting neurons, a majority (14/25 neurons) were inhibited by reward cues and activated by footshocks, while 7/25 neurons responded to only one of the two stimuli, with the remaining four neurons showing no response to either stimulus. Furthermore, individual neuron responses to reward cues correlated negatively with responses to footshocks (r2 = 0.2661, p = 0.0070) (Figure 4F). In marked contrast to VTA-projecting neurons, DRN-projecting neurons on average showed no responses to reward cues, although there was an average activation by footshocks, similar to VTA-projecting neurons (Figure 4E). Analysis of individual neurons further showed that responses to reward cues and footshocks did not correlate with each other (p = 0.9775) (Figure 4G). Overall, VTA- and DRN-projecting neurons exhibited markedly different proportions of neurons that were inhibited by reward cues (p = 0.001, Chi-square) (Figure 4H). Thus, our results indicate that VTA-projecting but not DRN-projecting RMTg neurons appear highly enriched in negative valence-encoding patterns. Excitotoxic RMTg lesions abolished VTA inhibitions by aversive stimuli Our findings that RMTg neurons are activated by aversive stimuli and VTA-projecting neurons are preferentially negative valence-encoding suggest that the RMTg could drive VTA responses to aversive stimuli, a hypothesis we tested by recording VTA neuron responses to phasic aversive stimuli with or without RMTg lesions (Figure 5E, Figure 5—figure supplement 1). NeuN staining showed that lesions were mostly restricted to the RMTg area and did not extend to surrounding structures such as the pedunculopontine (PPTg) and dorsal raphe nuclei (DRN) (Figure 5E). We classified recorded VTA neurons as being putative dopamine neurons (pDA neurons) by their phasic activations to reward cues (Sham: n = 21 pDA, n = 45 non-pDA, Lesion: n = 17 pDA, n = 40 non-pDA, p < 0.05 during 0–200 ms after reward cues), an activity signature repeatedly shown to correlate highly with optogenetically identified DA neurons in mice (Cohen et al., 2012; Eshel et al., 2015; Matsumoto et al., 2016) (Figure 5A,B). We also observed about 1/3 of recorded VTA neurons in sham lesioned rats exhibiting ramping activities in responses to reward cues (p < 0.05 during 200–2000 ms after reward cue onset), which are consistent with firing patterns found in genetically identified GABAergic neurons in the VTA (Cohen et al., 2012; Eshel et al., 2015) (Figure 5A,B). Additionally, 8 of the 66 recorded VTA neurons in sham group showed inhibitory responses to reward cues. Figure 5 with 2 supplements see all Download asset Open asset Excitotoxic RMTg lesions abolished VTA inhibitions by aversive stimuli. (A, B) Heatmap showing all VTA neuron responses to reward cues in sham group. pDA neurons in the VTA were classified by their phasic activation to reward cues (0–200 ms post-stimulus window), while pGABA neurons were classified by the presence of sustained activations (200–2000 ms post-stimulus window). (C) Raster plots of immediate inhibition and delayed inhibition response types observed in pDA neuron after aversive stimuli. (D) Comparisons of RMTg (blue trace) and pDA (red trace) neuron responses to affective stimuli. All three aversive stimuli elicited initial excitations in both RMTg and pDA neurons, after which RMTg neurons remained excited while pDA neurons showed inhibition during 100–500 ms window post-stimulus (brown-shaded boxes). pDA neurons were activated by reward cues, and at faster latencies than RMTg inhibition to the same cue, making it unlikely that pDA activations to the reward cue would be driven by the RMTg. (E) NeuN staining showed that lesions were mostly restricted to the RMTg area and did not extend to surrounding structures such as the pedunculopontine nucleus (PPTg) and dorsal raphe nucleus (DRN). Scalebars: 1 mm and 100 μm for left and right panels. (F) RMTg lesion (dashed trace) eliminated aversion-induced inhibition in pDA neurons. (G) Bar graphs again showing loss of aversion-induced inhibition during 100–500 ms in pDA neurons after RMTg lesions. (H) Loss of aversion-induced inhibition in all recorded VTA neurons. https://doi.org/10.7554/eLife.41542.008 We found that most pDA neurons in the sham group showed inhibitions to aversive stimuli, in some cases after a brief initial excitation (Figure 5C). Specifically, 86% of stimulus-responsive pDA neurons were inhibited by footshock (54% showing only inhibition; 32% showing delayed inhibition after a brief initial excitation), 75% by siren (28% showing only inhibition; 47% showing delayed inhibition), and 56% by bright light (27% showing only inhibition; 29% showing delayed inhibition), while the remaining showed excitation only (Figure 5—figure supplement 2A–C). Notably, inhibition by aversive stimuli was most prominent during a 100–500 ms window post-stimulus, while activation by reward cues was most prominent in an earlier 0–200 ms window post-stimulus, a timing pattern opposite to that of RMTg neurons in which responses to aversive stimuli tended to be faster by a few hundred milliseconds (Figure 5D). Among recorded VTA neurons, we found that RMTg lesions did not affect the magnitude nor the percentage of neurons responding to reward cues (0–200 ms post-stimulus) (p > 0.05, two-way ANOVA, Bonferroni test for multiple comparison, p = 0.879, Chi-square) (Figure 5—figure supplement 2D,E). Hence, we further analyzed reward-activated VTA neurons on the presumption that these were pDA in both lesioned and unlesioned animals. We found that, after RMTg lesions, pDA neurons no longer showed an average inhibition to any of the three phasic aversive stimuli (Figure 5F), and the proportions of pDA neurons inhibited by aversive stimuli were dramatically reduced to 19%, 16%, and 18% for footshock, siren, and bright light, respectively (p < 0.0001, p = 0.008, and p = 0.048, Chi-square, 100–500 ms post-stimulus) (Figure 5G). Due to the potential inaccuracy of our classification of pDA neurons, we then analyzed all recorded VTA neurons instead of pDA neurons alone. Consistent with pDA analysis, the proportions of inhibitory responses were also reduced by RMTg lesions, although to a less extreme degree than the pDA population. Specifically, in intact animals, 59%, 42%, and 36% of VTA neurons showed significant inhibitions to footshock, siren, and bright light, while in RMTg-lesioned rats, these proportions were reduced to 21%, 16%, and 9%, respectively (p = 0.005, p = 0.017, and p = 0.004, for footshock, siren, and bright light, respectively, Chi-square) (Figure 5H). While pDA inhibitory responses to aversive stimuli were reduced by RMTg lesions, the proportions of pDA neurons showing pure excitation were either marginally increased or unaffected (p = 0.182, p = 0.047, and p = 0.766 for siren, footshock, and bright light, respectively, Chi-square) (Figure 5—figure supplement 2G). As noted above, many pDA neurons that were inhibited by aversive stimuli did so after brief initial excitations, and RMTg lesions increased the magnitudes of these initial excitations relative to shams (p = 0.013, p = 0.021, and p = 0.009 for siren, shock, and bright light, respectively, unpaired t-test), without affecting the proportion of pDA neurons exhibiting them (p = 0.169, p = 0.328, and p = 0.126 for siren, shock, and bright light, respectively, Chi-square) (Figure 5—figure supplement 2H). Notably, RMTg lesions did not alter basal firing rates of pDA neurons (5.4 Hz vs 5.2 Hz, n = 23, p = 0.507, unpaired t-test). However, we did find that RMTg lesions increased the percentage of spikes found in bursts of pDA neurons (p = 0.014, unpaired t-test) (Figure 5—figure supplement 2F). Together, RMTg lesions influence both inhibitions and excitations of pDA neurons to aversive stimuli, greatly reducing inhibitions, and modestly increasing excitations to these stimuli, while having no apparent effect on responses to reward predictive cues nor neutral tones. RMTg lesions disrupt conditioned place aversion to a wide range of stimuli Finally, we examined the effect of RMTg lesions on conditioned place preference for three of the aversive stimuli that we tested earlier: siren, bright light, LiCl, and the delayed aversive phase of cocaine (Figure 6A, Figure 6—figure supplement 6–). Using a three-chambered apparatus, we were able to measure the effect of conditioning on both the time spent in each chamber (stimulus-paired, unpaired, and neutral) as well as the relative number of entries into the paired and unpaired chambers (Fi