The adenosine hypothesis of schizophrenia into its third decade: From neurochemical imbalance to early life etiological risks.

The adenosine hypothesis of schizophrenia was conceptualized about two decades ago in an attempt to integrate two prominent theories of neurochemical imbalance that attribute the pathogenesis of schizophrenia to hyperfunction of the mesocorticolimbic dopamine neurotransmission and hypofunction of cortical glutamate neurotransmission. Given its unique position as an endogenous modulator of both dopamine and glutamate signaling in the brain, adenosine was postulated as a potential new drug target to achieve multiple antipsychotic actions. This new strategy may offer hope for improving treatment, especially in alleviating negative symptoms and cognitive deficits of schizophrenia that do not respond to current medications. To date, however, the adenosine hypothesis has yet led to any significant therapeutic breakthroughs. Here, we address two possible reasons for the impasse. First, neither the presence of adenosine functional deficiency in people with schizophrenia nor its causal relationship to symptom production has been satisfactorily examined. Second, the lack of novel adenosine-based drugs also impedes progress. This review updates the latest preclinical and clinical data pertinent to the construct validity of the adenosine hypothesis and explores novel molecular processes whereby dysregulation of adenosine signaling could be linked to the etiology of schizophrenia. It is intended to stimulate and revitalize research into the adenosine hypothesis towards the development of a new and improved generation of antipsychotic drugs that has eluded us for decades.

Laboratory earthquake forecasting: A machine learning competition.

Earthquake prediction, the long-sought holy grail of earthquake science, continues to confound Earth scientists. Could we make advances by crowdsourcing, drawing from the vast knowledge and creativity of the machine learning (ML) community? We used Google's ML competition platform, Kaggle, to engage the worldwide ML community with a competition to develop and improve data analysis approaches on a forecasting problem that uses laboratory earthquake data. The competitors were tasked with predicting the time remaining before the next earthquake of successive laboratory quake events, based on only a small portion of the laboratory seismic data. The more than 4,500 participating teams created and shared more than 400 computer programs in openly accessible notebooks. Complementing the now well-known features of seismic data that map to fault criticality in the laboratory, the winning teams employed unexpected strategies based on rescaling failure times as a fraction of the seismic cycle and comparing input distribution of training and testing data. In addition to yielding scientific insights into fault processes in the laboratory and their relation with the evolution of the statistical properties of the associated seismic data, the competition serves as a pedagogical tool for teaching ML in geophysics. The approach may provide a model for other competitions in geosciences or other domains of study to help engage the ML community on problems of significance.

Homophily influences ranking of minorities in social networks.

Homophily can put minority groups at a disadvantage by restricting their ability to establish links with a majority group or to access novel information. Here, we show how this phenomenon can influence the ranking of minorities in examples of real-world networks with various levels of heterophily and homophily ranging from sexual contacts, dating contacts, scientific collaborations, and scientific citations. We devise a social network model with tunable homophily and group sizes, and demonstrate how the degree ranking of nodes from the minority group in a network is a function of (i) relative group sizes and (ii) the presence or absence of homophilic behaviour. We provide analytical insights on how the ranking of the minority can be improved to ensure the representativeness of the group and correct for potential biases. Our work presents a foundation for assessing the impact of homophilic and heterophilic behaviour on minorities in social networks.

Negative transfer effects between reference memory and working memory training in the water maze in C57BL/6 mice.

The water maze is one of the most widely employed spatial learning paradigms in the cognitive profiling of genetically modified mice. Oftentimes, tests of reference memory (RM) and working memory (WM) in the water maze are sequentially evaluated in the same animals. However, critical difference in the rules governing efficient escape from the water between WM and RM tests is expected to promote the adoption of incompatible mnemonic or navigational strategies. Hence, performance in a given test is likely poorer if it follows the other test instead of being conducted first. Yet, the presence of such negative transfer effects (or proactive interference) between WM and RM training in the water maze is often overlooked in the literature. To gauge whether this constitutes a serious concern, the present study determined empirically the magnitude, persistence, and directionality of the transfer effect in wild-type C57BL/6 mice. We contrasted the order of tests between two cohorts of mice. Performance between the two cohorts in the WM and RM tests were then separately compared. We showed that prior training of either test significantly reduced performance in the subsequent one. The statistical effect sizes in both directions were moderate to large. Although extended training could overcome the deficit, it could re-emerge later albeit in a more transient fashion. Whenever RM and WM water maze tests are conducted sequentially in the same animals - regardless of the test order, extra caution is necessary when interpreting the outcomes in the second test. Counterbalancing test orders between animals is recommended.

JANUS: A hypothesis-driven Bayesian approach for understanding edge formation in attributed multigraphs.

Understanding edge formation represents a key question in network analysis. Various approaches have been postulated across disciplines ranging from network growth models to statistical (regression) methods. In this work, we extend this existing arsenal of methods with JANUS, a hypothesis-driven Bayesian approach that allows to intuitively compare hypotheses about edge formation in multigraphs. We model the multiplicity of edges using a simple categorical model and propose to express hypotheses as priors encoding our belief about parameters. Using Bayesian model comparison techniques, we compare the relative plausibility of hypotheses which might be motivated by previous theories about edge formation based on popularity or similarity. We demonstrate the utility of our approach on synthetic and empirical data. JANUS is relevant for researchers interested in studying mechanisms explaining edge formation in networks from both empirical and methodological perspectives.

Evidence of Online Performance Deterioration in User Sessions on Reddit.

This article presents evidence of performance deterioration in online user sessions quantified by studying a massive dataset containing over 55 million comments posted on Reddit in April 2015. After segmenting the sessions (i.e., periods of activity without a prolonged break) depending on their intensity (i.e., how many posts users produced during sessions), we observe a general decrease in the quality of comments produced by users over the course of sessions. We propose mixed-effects models that capture the impact of session intensity on comments, including their length, quality, and the responses they generate from the community. Our findings suggest performance deterioration: Sessions of increasing intensity are associated with the production of shorter, progressively less complex comments, which receive declining quality scores (as rated by other users), and are less and less engaging (i.e., they attract fewer responses). Our contribution evokes a connection between cognitive and attention dynamics and the usage of online social peer production platforms, specifically the effects of deterioration of user performance.

Inhibition of glycine transporter 1: The yellow brick road to new schizophrenia therapy?

While pharmacological blockade of dopamine D2 receptor can effectively suppress the psychotic or positive symptoms of schizophrenia, there is no satisfactory medication for the negative and cognitive symptoms of schizophrenia in spite of the proliferation of second generation antipsychotic drugs. Excitements over a new class of third generation antipsychotics that might possibly fill this urgent medical need have been prompted by the recent development of glycine transporter 1 (GlyT1) inhibitors. The impetus of this novel pharmacological strategy stems directly from the prevailing hypothesis that negative and cognitive symptoms are attributable to the hypofunction of glutamatergic signalling via the N-methyl-D-aspartate (NMDA) receptor in the brain. Inhibition of GlyT1 reduces clearance of extra-cellular glycine near NMDA receptor-containing synapses, and thereby increases baseline occupancy of the glycine-B site at the NR1 subunit of the NMDA receptor, which is a prerequisite of channel activation upon stimulation by the excitatory neurotransmitter glutamate. Pharmacological inhibition of GlyT1 is expected to boost NMDA receptor function and therefore alleviate persistent negative and cognitive symptoms without excessive risk of excitotoxicity associated with direct NMDA receptor agonists. The recently completed phase III clinical trials of the Roche compound, bitopertin (a.k.a. RG1678 or RO-4917838) had initially raised hope that this new class of drugs might represent the first successful translation of the glutamate hypothesis of schizophrenia to the clinic. However, the outcomes of the multi-centre bitopertin clinical trials have been disappointing. The present review seeks to examine this promise through a critical survey of the latest clinical and preclinical findings on the therapeutic potential of GlyT1 inhibition or down-regulation.

Caffeine impairs the acquisition and retention, but not the consolidation of Pavlovian conditioned freezing in mice.

RATIONALE: The psychoactive substance, caffeine, may improve cognitive performance, but its direct impact on learning and memory remains ill defined. Conflicting reports suggest that caffeine may impair as well as enhance Pavlovian fear conditioning in animals and its effect may vary across different phases of learning. OBJECTIVES: The purpose of this study is to dissect the effect of a motor-stimulant dose of caffeine (30 mg/kg intraperitoneal (i.p.)) on acquisition, retrieval or consolidation of conditioned fear in C57BL/6 mice. METHODS: Fear conditioning was evaluated in a conditioned freezing paradigm comprising 3 tone-shock pairings and a two-way active avoidance paradigm lasting two consecutive days with 80 conditioning trials per test session. RESULTS: Conditioning to both the discrete tone-conditioned stimulus (CS) and the context was markedly impaired by caffeine. The deficits were similarly evident when caffeine was administered prior to acquisition or retrieval (48 and 72 h after conditioning); and the most severe impairment was seen in animals given caffeine before acquisition and before retrieval. A comparable deficit was observed in the conditioned active avoidance test. By contrast, caffeine administered immediately following acquisition neither affected the expression of tone freezing nor context freezing. CONCLUSIONS: The present study challenges the previous report that caffeine primarily disrupts hippocampus-dependent conditioning to the context. At the relevant dose range, acute caffeine likely exerts more widespread impacts beyond the hippocampus, including the amygdala and striatum that are anatomically connected to the hippocampus; together, they support the acquisition and retention of fear memories to discrete stimuli as well as diffused contextual cues.

Deletion of forebrain glycine transporter 1 enhances conditioned freezing to a reliable, but not an ambiguous, cue for threat in a conditioned freezing paradigm.

Enhanced expression of Pavlovian aversive conditioning but not appetitive conditioning may indicate a bias in the processing of threatening or fearful events. Mice with disruption of glycine transporter 1 (GlyT1) in forebrain neurons exhibit such a bias, but they are at the same time highly sensitive to manipulations that hinder the development of the conditioned response (CR) suggesting that the mutation may modify higher cognitive processes that extract predictive information between environmental cues. Here, we further investigated the development of fear conditioning in forebrain neuronal GlyT1 knockout mice when the predictiveness of a tone stimulus for foot-shock was rendered ambiguous by interspersing [tone→no shock] trials in-between [tone→shock] trials during acquisition. The CR to the ambiguous tone CS (conditioned stimulus) was compared with that generated by an unambiguous CS that was always followed by the shock US (unconditioned stimulus) during acquisition. We showed that rendering the CS ambiguous as described significantly attenuated the CR in the mutants, but it was not sufficient to modify the CR in the control mice. It is concluded that disruption of GlyT1 in forebrain neurons does not increase the risk of forming spurious and potentially maladaptive fear associations.

Detecting memory and structure in human navigation patterns using Markov chain models of varying order.

One of the most frequently used models for understanding human navigation on the Web is the Markov chain model, where Web pages are represented as states and hyperlinks as probabilities of navigating from one page to another. Predominantly, human navigation on the Web has been thought to satisfy the memoryless Markov property stating that the next page a user visits only depends on her current page and not on previously visited ones. This idea has found its way in numerous applications such as Google's PageRank algorithm and others. Recently, new studies suggested that human navigation may better be modeled using higher order Markov chain models, i.e., the next page depends on a longer history of past clicks. Yet, this finding is preliminary and does not account for the higher complexity of higher order Markov chain models which is why the memoryless model is still widely used. In this work we thoroughly present a diverse array of advanced inference methods for determining the appropriate Markov chain order. We highlight strengths and weaknesses of each method and apply them for investigating memory and structure of human navigation on the Web. Our experiments reveal that the complexity of higher order models grows faster than their utility, and thus we confirm that the memoryless model represents a quite practical model for human navigation on a page level. However, when we expand our analysis to a topical level, where we abstract away from specific page transitions to transitions between topics, we find that the memoryless assumption is violated and specific regularities can be observed. We report results from experiments with two types of navigational datasets (goal-oriented vs. free form) and observe interesting structural differences that make a strong argument for more contextual studies of human navigation in future work.

Discovering beaten paths in collaborative ontology-engineering projects using Markov chains.

Biomedical taxonomies, thesauri and ontologies in the form of the International Classification of Diseases as a taxonomy or the National Cancer Institute Thesaurus as an OWL-based ontology, play a critical role in acquiring, representing and processing information about human health. With increasing adoption and relevance, biomedical ontologies have also significantly increased in size. For example, the 11th revision of the International Classification of Diseases, which is currently under active development by the World Health Organization contains nearly 50,000 classes representing a vast variety of different diseases and causes of death. This evolution in terms of size was accompanied by an evolution in the way ontologies are engineered. Because no single individual has the expertise to develop such large-scale ontologies, ontology-engineering projects have evolved from small-scale efforts involving just a few domain experts to large-scale projects that require effective collaboration between dozens or even hundreds of experts, practitioners and other stakeholders. Understanding the way these different stakeholders collaborate will enable us to improve editing environments that support such collaborations. In this paper, we uncover how large ontology-engineering projects, such as the International Classification of Diseases in its 11th revision, unfold by analyzing usage logs of five different biomedical ontology-engineering projects of varying sizes and scopes using Markov chains. We discover intriguing interaction patterns (e.g., which properties users frequently change after specific given ones) that suggest that large collaborative ontology-engineering projects are governed by a few general principles that determine and drive development. From our analysis, we identify commonalities and differences between different projects that have implications for project managers, ontology editors, developers and contributors working on collaborative ontology-engineering projects and tools in the biomedical domain.

Sensorimotor gating is disrupted by acute but not chronic systemic exposure to caffeine in mice.

RATIONALE: Caffeine is a psychostimulant drug that blocks adenosine A1 and A2A receptors (A1Rs and A2ARs). However, its ability to disrupt early sensory gating as indexed by prepulse inhibition (PPI), which is consistently disrupted by other psychostimulant agents, has never been convincingly demonstrated. OBJECTIVES: To compare the impact of caffeine on PPI expression in C57BL/6 mice by two dose-response experiments differing in terms of chronicity, regimen, and route of administration. To study separately the acute effect of selective antagonists against A1R or A2AR. METHODS: Caffeine (10, 30, 100 mg/kg, intraperitoneal (i.p.)) was either administered shortly before testing or via caffeinated drinking water (0.3, 1.0, 2 g/l) in home cages over 3 weeks. Two separate dose-response studies tested the acute effect of the selective A1R antagonist, 1,3 dipropyl-8 cyclopentyl xanthine (DPCPX), and the selective A2AR antagonist, 5-amino-7-(2-phenylethyl)-2-(2-furyl)-pyrazolo-[4,3-e]-1,2,4-triazolo[1,5-c] (SCH 58261) (0.2, 1.0, 5.0 mg/kg, i.p.). The two drugs were combined in a final experiment to identify their potential synergistic interaction. RESULTS: While the two lower acute doses of caffeine attenuated PPI, the highest dose potentiated PPI. By contrast, chronic caffeine exposure did not affect PPI. Neither DPCPX nor SCH 58261 altered PPI, and no synergism was observed when the two drugs were combined. CONCLUSIONS: This is the first demonstration that acute caffeine disrupts PPI, but the relative contribution of A1R and A2AR blockade remains unclear, and possible non-adenosinergic mechanisms cannot be ruled out. The null effect under chronic caffeine exposure might involve the development of tolerance, but the precise receptor subtypes involved also warrant further investigation.

Forebrain glycine transporter 1 deletion enhances sensitivity to CS-US discontiguity in classical conditioning.

The deletion of glycine transporter 1 (GlyT1) in forebrain neurons can apparently strengthen Pavlovian aversive conditioning, but this phenotype is not expressed if conditioning followed non-reinforced pre-exposures of the to-be-conditioned stimulus (CS). To examine whether GlyT1 disruption may only enhance aversive associative learning under conditions that most favour the formation of CS-US excitatory link, we evaluated the impact of GlyT1 disruption on the trace conditioning procedure whereby a trace interval between a tone CS and a shock US was introduced during conditioning. CS and US occurrences were thus rendered discontiguous, which was expected to impede conditioning compared with contiguous CS-US pairing. Conditioned freezing to the CS was measured in a retention test conducted 48 h after conditioning. The genetic disruption significantly modified the temporal dynamics of the freezing response over the course of the 8-min presentation of the CS, although the immediate conditioned response to the CS was unaffected. The separation between "trace" and "no-trace" conditions was augmented in the mutant mice, but this only became apparent in mid-session; and the augmentation can be attributed to the combined effects of (i) weaker conditioned freezing in the mutant relative to control subjects in the "trace" condition, and (ii) stronger conditioned freezing in mutants relative controls in the "no-trace" condition. The demonstrated increased sensitivity to the effect of CS-US temporal discontiguity further highlights the importance of GlyT1-dependent mechanisms in the regulation of associative learning.

Regulation of fear responses by striatal and extrastriatal adenosine A2A receptors in forebrain.

BACKGROUND: Adenosine A2A receptors (A2ARs) are enriched in the striatum but are also present at lower levels in the extrastriatal forebrain (i.e., hippocampus, cortex), integrating dopamine, glutamate, and brain-derived neurotrophic factor (BDNF) signaling, and are thus essential for striatal neuroplasticity and fear and anxiety behavior. METHODS: We tested two brain region-specific A2AR knockout lines with A2ARs selectively deleted either in the striatum (st-A2AR KO) or the entire forebrain (striatum, hippocampus, and cortex [fb-A2AR KO]) on fear and anxiety-related responses. We also examined the effect of hippocampus-specific A2AR deletion by local injection of adeno-associated virus type 5 (AAV5)-Cre into floxed-A2AR knockout mice. RESULTS: Selectively deleting A2ARs in the striatum increased Pavlovian fear conditioning (both context and tone) in st-A2AR KO mice, but extending the deletion to the rest of the forebrain apparently spared context fear conditioning and attenuated tone fear conditioning in fb-A2AR KO mice. Moreover, focal deletion of hippocampal A2ARs by AAV5-Cre injection selectively attenuated context (but not tone) fear conditioning. Deletion of A2ARs in the entire forebrain in fb-A2AR KO mice also produced an anxiolytic phenotype in both the elevated plus maze and open field tests, and increased the startle response. These extrastriatal forebrain A2AR behavioral effects were associated with reduced BDNF levels in the fb-A2AR KO hippocampus. CONCLUSIONS: This study provides evidence that inactivation of striatal A2ARs facilitates Pavlovian fear conditioning, while inactivation of extrastriatal A2ARs in the forebrain inhibits fear conditioning and also affects anxiety-related behavior.

SSR504734 enhances basal expression of prepulse inhibition but exacerbates the disruption of prepulse inhibition by apomorphine.

RATIONALE: Inhibition of glycine transporter 1 (GlyT1) elevates extracellular glycine and can thus increase N-methyl-D-aspartate receptor (NMDAR) excitability in the brain. The potent GlyT1 inhibitor, SSR504734, has also been shown to potentiate the behavioral effects of direct and indirect dopamine agonists. Thus, an acute systemic dose of SSR504734 was sufficient to exacerbate the motor-stimulant effect of the dopamine releaser amphetamine in C57BL/6 mice, even though SSR504734 alone exerted no significant effect on motor activity. OBJECTIVES: Here, we explore if SSR504734 might modulate dopamine-dependent sensory gating in the paradigm of prepulse inhibition (PPI) of the acoustic startle reflex. METHODS: Experiment 1 characterized the effect of SSR504734 (10 and 30 mg/kg i.p.) on PPI expression when administered alone. Experiments 2 and 3 investigated the impact of SSR504734 when administered in conjunction with the dopamine receptor agonist, apomorphine (1 and 2 mg/kg s.c.), which is known to reliably disrupt PPI. RESULTS: When administered alone, acute SSR504734 enhanced PPI only at 30 mg/kg--a dose that has been shown to improve cognitive functions including working memory, which has been linked to enhanced NMDAR function resulting from the elevation of extracellular glycine. However, this effect did not allow SSR504734 to antagonize the PPI-disruptive effect of apomorphine. At the lower dose of 10 mg/kg--that was insufficient to enhance PPI when administered alone--SSR504734 even exacerbated the deleterious effect of apomorphine on PPI. CONCLUSIONS: The therapeutic potential of GlyT1 inhibition against distinct behavioral/cognitive deficiency might require different magnitudes of GlyT1 inhibition.

A conceptual and practical guide to the behavioural evaluation of animal models of the symptomatology and therapy of schizophrenia.

Schizophrenia is a chronic debilitating brain disorder characterized by a complex set of perceptual and behavioural symptoms that severely disrupt and undermine the patient's psychological well-being and quality of life. Since the exact disease mechanisms remain essentially unknown, holistic animal models are indispensable tools for any serious investigation into the neurobiology of schizophrenia, including the search for remedies, prevention of the disease and possible biological markers. This review provides some practical advice to those confronted with the task of evaluating their animal models for relevance to schizophrenia, a task that inevitably involves behavioural tests with animals. To a novice, this challenge not only is a technical one but also entails attention to interpretative issues concerning validity and translational power. Here, we attempt to offer some guidance to help overcome these obstacles by drawing on our experience of diverse animal models of schizophrenia based on genetics, strain difference, brain lesions, pharmacological induction and early life developmental manipulations. The review pays equal emphasis to the general (theoretical) considerations of experimental design and the illustration of the problems related to critical test parameters and the data analysis of selected exemplar behavioural tests. Finally, the individual differences of behavioural expression in relevant tests observed in wild-type animals might offer an alternative approach in order to explore the mechanism of schizophrenia-related behavioural dysfunction at the molecular, cellular and structural levels, all of which are of more immediate relevance to cell and tissue research.

Prepulse inhibition predicts working memory performance whilst startle habituation predicts spatial reference memory retention in C57BL/6 mice.

Prepulse inhibition (PPI) of the acoustic startle reflex refers to the attenuation of the startle response to an intense pulse stimulus when it is shortly preceded by a weak non-startling prepulse stimulus. It is a well-established high-throughput translational measure of pre-attentive sensory gating, and its impairment is detected in several neuropsychiatric diseases including schizophrenia. It has been hypothesized that PPI might be associated with, or predictive of, cognitive deficiency in such diseases, and therefore provide an efficient assay for screening drugs with potential pro-cognitive efficacy. Free from any predetermined disease model, the present study evaluated in a homogeneous cohort of inbred C57BL/6 mice the presence of a statistical link between PPI expression and cognitive performance. Performance indices in a spatial reference memory test and a working memory test conducted in the Morris water maze, and contextual fear conditioning were correlated against pre-existing baseline PPI expression. A specific correlative link between working memory and PPI induced by weak (but not strong) prepulse was revealed. In addition, a correlation between habituation of the startle reflex and reference memory was identified for the first time: a stronger overt habituation effect was associated with superior spatial search accuracy. The PPI paradigm thus provides two independent predictors of dissociable cognitive traits in normal C57BL/6 mice; and they might serve as potential markers for high-throughput evaluation of potential cognitive enhancers, especially in the context of schizophrenia where deficits in startle habituation and PPI co-exist.

Deletion of striatal adenosine A(2A) receptor spares latent inhibition and prepulse inhibition but impairs active avoidance learning.

Following early clinical leads, the adenosine A(2A)R receptor (A(2A)R) has continued to attract attention as a potential novel target for treating schizophrenia, especially against the negative and cognitive symptoms of the disease because of A(2A)R's unique modulatory action over glutamatergic in addition to dopaminergic signaling. Through (i) the antagonistic interaction with the dopamine D(2) receptor, and (ii) the regulation of glutamate release and N-methyl-d-aspartate receptor function, striatal A(2A)R is ideally positioned to fine-tune the dopamine-glutamate balance, the disturbance of which is implicated in the pathophysiology of schizophrenia. However, the precise function of striatal A(2A)Rs in the regulation of schizophrenia-relevant behavior is poorly understood. Here, we tested the impact of conditional striatum-specific A(2A)R knockout (st-A(2A)R-KO) on latent inhibition (LI) and prepulse inhibition (PPI) - behavior that is tightly regulated by striatal dopamine and glutamate. These are two common cross-species translational tests for the assessment of selective attention and sensorimotor gating deficits reported in schizophrenia patients; and enhanced performance in these tests is associated with antipsychotic drug action. We found that neither LI nor PPI was significantly affected in st-A(2A)R-KO mice, although a deficit in active avoidance learning was identified in these animals. The latter phenotype, however, was not replicated in another form of aversive conditioning - namely, conditioned taste aversion. Hence, the present study shows that neither learned inattention (as measured by LI) nor sensory gating (as indexed by PPI) requires the integrity of striatal A(2A)Rs - a finding that may undermine the hypothesized importance of A(2A)R in the genesis and/or treatment of schizophrenia.

Dysregulation of brain adenosine is detrimental to the expression of conditioned freezing but not general Pavlovian learning.

Glutamatergic and dopaminergic neurotransmission is modulated by adenosine, whose ambient level in the brain is in turn regulated by the metabolic enzyme, adenosine kinase (ADK). Brain adenosinergic tone can therefore be effectively reduced and increased by up- and down-regulation of ADK expression, respectively. Although changes in brain ADK levels can yield multiple behavioral effects, the precise functional significance of telencephalon (neocortical and limbic structures) adenosine remains ill-defined. Among the phenotypes identified in transgenic mice with brain-wide ADK overexpression (ADK(TG) mice) and reduced adenosinergic tone, working memory deficiency and potentiated response to systemic N-methyl-d-aspartate receptor blockade were exacerbated by the introduction of local ADK disruption (elevated adenosinergic tone) restricted to the telencephalon (ADK(TG):ADK(Tel-def) mice). These two phenotypes, which are central to schizophrenia cognitive/negative symptoms, appear to be regulated by adenosinergic activities within and outside the telencephalon in a complementary manner. Here, we extended this unique comparison between ADK(TG) mice ADK(TG):ADK(Tel-def) mice to another prominent phenotype previously documented in ADK(TG) mice - namely, impaired Pavlovian conditioned freezing. We found that ADK(TG):ADK(Tel-def) mice again were associated with a more severe phenotype while sharing a similar phenotype profile. Furthermore, we qualified that this Pavlovian phenotype did not translate into a general deficiency in associative learning, since no such deficit was evident in three other (aversive and appetitive) Pavlovian learning paradigms. The present study has thus identified a hitherto unknown function of brain adenosine: the execution of conditioned freezing behavior, which is dependent on the balance of adenosinergic changes between the telencephalon and the rest of the brain.