A pharmacological investigation of behavioural flexibility in zebrafish (Danio rerio)

Abstract

Human and animal behaviour is ultimately directed at acquiring certain outcomes or achieving certain goals, or alternatively avoiding or mitigating negative outcomes. An intricate interplay between two components of action-outcome processing, namely goal-directed and habitual behaviour constantly surveys and manages behaviour to promote superior outcomes. Goal-directed behaviour is usually employed when behavioural responses are directed towards the completion of a specific, often novel outcome requiring considerable levels of cognitive deliberation. After frequent repetition of said behavioural actions, habitual behaviour begins to develop with the apparent objective of saving cognitive effort. This implies that not all habitual behaviour is entirely without ‘goal' since such outcomes normally retain functional value and desirability. However, once the subject demonstrates overreliance on habitual actions, even when the behaviour persists in the absence of a specific functional outcome, psychopathologies, e.g. obsessive-compulsive disorder (OCD)1 are seen. This interplay between goal-directed and habitual behavioural selection is constantly modulated on a neurocognitive continuum, with cognitive flexibility (CF)2 on its one end, and cognitive rigidity (CR)3, on the other. CF can loosely be defined as the ability to appropriately adjust one’s behaviour in response to a changing environment, inter alia allowing for the selective engagement in tasks of a higher immediate priority. CR, on the other hand, is characterised by slow or even inadequate responses to changing outcomes or a changing environment. As such, when habitual behavioural engagement becomes excessive, an imbalance has formed within the neurocognitive continuum and CR begins to overwhelm CF. Habitual and goal-directed task execution, hence also cognitive action-outcome processing, is broadly founded upon cortico-striatal-thalamic-cortical (CSTC)4 signalling at neuroanatomical level. Although this system is differentially modulated by various neurotransmitters, dopamine seems to be of utmost importance. Specifically, it is proposed that excessive dopaminergic signalling, either directly or indirectly via mechanisms of inadequate serotonergic control, is believed to underlie hyperactive CSTC processing. Excessive dopaminergic signalling modulates behaviour by acting on two differentially identified neural pathways in the CSTC circuit, namely the behaviourally activating direct pathway and the behaviourally inactivating indirect pathway. It is further known that excessive dopaminergic activity results in the expression of persistent and repetitive behavioural phenotypes that collectively represent manifestations of CR. Indeed, such an imbalance between CF and CR, and by extension abnormal activity in CSTC circuit, has been demonstrated in a number of psychiatric disorders, e.g. OCD, major depressive

disorder (MDD)1, and autism spectrum disorder (ASD)2. That said, most of these disorders show a suboptimal treatment response to current treatment options, highlighting a need for alternative approaches which may yield superior results. Therefore, it is believed that targeting a common neuropsychological construct, i.e. CR3, that underlie said conditions, might lead to the development of more promising treatment avenues. One potential method to pharmacologically interrogate a construct like CR would be to use a reward-feedback learning paradigm, which is often exploited in psychology in the form of conditioned learning, e.g. classic Pavlovian conditioning. Briefly, in a cue-reward contingency perspective, a subject is trained that a specific sensory cue is associated with a tangible rewarding outcome, and thus the presentation of the cue begins to stimulate approaching behaviour. Thereafter, once the contingency has been firmly acquired, but then changed so that the cue no longer predicts the experience of the reward, processes of reversal learning are enacted. Under normal, healthy circumstances, the subject is then able to disengage from the once relevant cue and alter their behaviour accordingly since the presentation of the cue becomes irrelevant to the experience of the reward. However, individuals suffering from CR, seem to be quite resistant to such changes. Therefore, the aim of the current investigation was to build on previous work done in our laboratory which utilized zebrafish in a cue-reward contingency learning paradigm. In this context, the reward comprised a visual social reward, which was differentially presented with a cue in the form of a monochromatic pattern. Applying this, we aimed to explore the effect of chronic administration of the D1/2 receptor agonist, apomorphine (APO)4, on cue- or reward-directed behaviours and behavioural persistence. We further aimed to establish whether APO-associated behavioural changes would be modified by chronic exposure to levetiracetam (LEV)5, a potentially novel cognitive enhancer with a mechanism of action that is potentially selective for abnormally active neurons and pathways. The results obtained from this study confirmed the previous findings that showed zebrafish to be a suitable model system for investigations of reward-directed behavioural responses. Furthermore, we found that long-term dopaminergic potentiation suppressed reversal learning after fish acquired knowledge of a cue-reward contingency. However, with regards to our working hypothesis that LEV, being a putative cognitive enhancer that stabilizes excessive neuronal firing, would attenuate the behavioural effects elicited by APO, our results revealed the opposite. It is likely that the observed effect

of LEV1 in the present work, i.e. in blunting reward-directed responses, could be ascribed to the ability of LEV to blunt α-amino-3-hydroxyl-5-methyl-4-isoxazolepropionate (AMPA)2 receptor specific glutamatergic signalling, which has previously been shown to compromise cue-reward learning. Therefore, future studies are needed to explore this theory.