Assessment of intradimensional/extradimensional attentional set-shifting in rats

https://doi.org/10.1016/j.neubiorev.2018.02.013Get rights and content

Highlights

  • Rats learn novel discriminations based on odours and digging media equally well.

  • Reversal learning is more difficult when attending to odour.

  • Group-level attentional set-shifting performance is consistent between tests.

  • Individual-level attentional set-shifting performance is not consistent between tests.

  • Rats do not have a ‘flexibility quotient’.

Abstract

The rat intradimensional/extradimensional (ID/ED) task, first described by Birrell and Brown 18 years ago, has become the predominant means by which attentional set-shifting is investigated in rodents: the use of rats in the task has been described in over 135 publications by researchers from nearly 90 universities and pharmaceutical companies. There is variation in the protocols used by different groups, including differences in apparatus, stimuli (both stimulus dimensions and exemplars within), and also the methodology. Nevertheless, most of these variations seem to be of little consequence: there is remarkable similarity in the profile of published data, with consistency of learning rates and in the size and reliability of the set-shifting and reversal ‘costs’. However, we suspect that there may be inconsistent data that is unpublished or perhaps ‘failed experiments’ that may have been caused by unintended deviations from effective protocols. The purpose of this review is to describe our approach and the rationale behind certain aspects of the protocol, including common pitfalls that are encountered when establishing an effective local protocol.

Introduction

Rats and mice account for more than 70% of animals used in the UK under the Animals (Scientific Procedures) Act 1986, with a third of these being in the translational research category of ‘Applied – human medicine’ (UK Home Office, 2017). Although there has been a recent retreat from translational neuroscience in psychiatry, in part due to a lack of understanding the neurobiology of psychiatric disorders (Insel et al., 2012), research with non-human animals is providing important insight into the nature of cognitive impairments in conditions such as depression, dementia and psychosis. All of these conditions have impairments of so-called ‘executive functions’ of the frontal lobes, the severity of which are associated with poor functional outcome. Cognitive flexibility – “the ability to switch thought and/or response patterns” (Powell and Ragozzino, 2017) – is one such function: how the brain solves the problem of being, simultaneously, consistent and efficient (able to learn and generalise that learning to new situations) and yet also flexible (able to know that ‘things change’ and that ‘rules have exceptions’).

The early psychology literature is replete with a great variety of demonstrations of cognitive flexibility in many different contexts and in many species, ranging from fish to rodents and humans. Reversal learning has been called a “pre-eminent test of cognitive flexibility” (Izquierdo et al., 2017), not least because it is observed ubiquitously and is also easily quantified in different species. Other demonstrations of cognitive flexibility include task switching (Jersild, 1927), when response strategies need to change, and the shifting of attention as the relevance of perceptual features changes (demonstrated, for example, in the Wisconsin Card Sorting Test (Berg, 1948) and the intra/extradimensional (ID/ED) attentional set-shifting task (Lawrence, 1949)).

We have previously argued that task switching, attentional shifting and reversal learning are unlikely to reflect a unitary function called ‘cognitive flexibility’ (Brown and Tait, 2015). Shifting and switching tasks have in common the idea that prior experience causes the cognitive system to be dynamically set, or prepared, to perform particular mental operations or process particular information. This cognitive preparedness – also known as ‘mental set’ – confers a processing advantage (either stimulus processing in the case of a perceptual attentional set, or response selection in the case of a task or learning set) for as long as the preparation is appropriate. When the set of the system is not appropriate, the model-based processing will be disadvantageous, thus the system must be flexible and able to reset. The ID/ED task (Lawrence, 1949) enables this to be demonstrated by comparing new learning in two different states of mental set. At the ID stage, novel stimuli are presented but prior experience of particular perceptual features being relevant (e.g., colour) ensures that the processing of those features are prioritised, which confers an advantage for learning. At the ED stage, different perceptual features of novel stimuli (e.g., shape) are now relevant to solve the task, but as they are not the features prioritised, this results in a learning decrement. A comparison of learning rates in these two different states thus provides inference of the state of the mental set.

It is possible that the process or mechanisms that enable reversal learning may have been repurposed to support cognitive flexibility. In other words, cognitive (covert) flexibility could be a special case of behavioural (overt) flexibility. On the other hand, it seems more likely that reversal learning – like any learning – can occur in the context of various states of cognitive preparedness, ranging from model-free (no prior set) to entirely model-based, and this will probably be determined by the context or task variant (Izquierdo et al., 2017). ‘Learning set’ (Harlow, 1949) describes an increase in the rate of reversal learning as a function of experience of learning reversals, and it indicates that mental set (and its corollary, cognitive flexibility) is not an intrinsic, let alone necessary, aspect of adaptive behaviour resulting from learning processes (which includes reversal learning), but rather is additional to it. In other words, having a mental set (a model) can influence the rate of any learning, including reversal learning, but the nature of the mental set cannot be known by observing an isolated instance of learning. The mental set is only revealed by assessing the relative advantage or disadvantage that the model confers. This is one of the reasons we suggested that it is important that a task does not conflate reversal learning with either switching or shifting (Brown and Tait, 2015). This is particularly problematic in rule- or strategy-switching tasks for rats that employ mazes or operant chambers (see Floresco and Jentsch, 2011) because the responses to the different rules are not unique. On 50% of trials, the response to a new rule (e.g., “turn left”) will be the same as when an old rule (e.g., “approach the light”) is applied. This partial reinforcement effect, which is the result of a learning process, cannot be distinguished from the effects of cognitive flexibility. In shifting tasks, this problem can be overcome by having a sufficiently large number of stimulus exemplars so that it is possible to have a ‘total change design’: previously rewarded stimuli are no longer present and therefore not partially reinforced (Slamecka, 1968).

In summary, although we acknowledge that aspects of cognitive flexibility are undoubtedly relevant to, and can be assessed in the context of, reversal learning (see also Dhawan et al., in press), we do not think that all examples of reversal learning are relevant or that it is a simple way to measure cognitively flexibility. We think it is yet to be determined whether shifting and switching represent a unitary executive function, although the involvement of prefrontal cortex in both seems a compelling reason to suggest that these behaviours have aspects in common. The purpose of this paper is to describe our methods and protocol for assessing cognitive flexibility and our rationale for these. We do not intend to imply that we think this is the only, or even the best, way to assess these psychological constructs. Rather, we hope to provide helpful information for other researchers’ who might consider adopting or adapting the ID/ED attentional set-shifting task (ASST) for the rat.

The ID/ED ASST is a well-established behavioural assay which is used in humans, primates and rodents (for review see Brown and Tait, 2016). Performance in this task specifically is impaired in neurodegenerative diseases (e.g., Parkinson's disease; Downes et al., 1989) and neurological disorders (e.g., schizophrenia; Elliott et al., 1995) with frontocortical neuropathology, and in rodent models of these disorders (e.g., subchronic phencyclidine as a model of schizophrenia; Rodefer et al., 2005). We believe that the particular value of the task is that, regardless of species, the ID/ED ASST is formally the same: it requires the participant/subject to learn a series of two-choice compound discriminations with (typically) two systematically varied, uncorrelated stimulus dimensions – one is relevant to solving the discrimination (i.e., predicts reward), the other is irrelevant. Over multiple ASST stages, an attentional set is formed to the persistently relevant dimension, and then the participant/subjects’ ability to flexibly shift attention from that dimension to the previously irrelevant dimension is tested. The trials required to learn the discrimination at the ED stage is compared to learning at the ID stage and the difference is assumed to reflect the strength of the set and the cognitive cost (‘shift-cost’) of flexibility. Manipulations that increase shift-cost relative to control performance (which is generally expressed as additional trials at the ED stage, because there is often little room for improvement in ID acquisition) are typically interpreted as reflecting an impairment in cognitive flexibility, although the specific latent mechanisms can only be inferred. A reduced shift-cost is more difficult to interpret, as it could result from performance change at either ID (increased trials) or ED (decreased trials) or both (changes to shift-cost are discussed in more detail in Section 4.2.2).

An ID/ED ASST that is suitable for testing humans or monkeys typically uses compound (multidimensional) visual stimuli presented on a computerised touchscreen (Roberts et al., 1988). For example, the ID/ED ASST in the Cambridge Neuropsychological Test Automated Battery (CANTAB) (Cambridge Cognition, Ltd) uses stimuli which are opaque shapes with superimposed line-configurations. An ID/ED ASST employing a total-change design suitable for testing rodents was described by Birrell and Brown (2000). This approach relies on the natural propensity of rats and mice to forage for food, with subjects digging for food bait in small bowls which are discriminable by the digging media, or the scent, or the bowl itself may even have a different appearance or texture. This adaptation of the ASST for rodents, allows researchers to understand the same mechanisms governing attentional set-shifting in mammals, but using species-appropriate stimuli and responses. There is a common standard in the stages of the rat ASST: the majority of published designs use seven stages (Tait et al., 2014) – a simple discrimination (SD); a compound discrimination (CD); a reversal of the CD (REV1); the ID; a reversal of the ID (REV2); the ED; and finally a reversal of the ED (REV3) – which we refer to as the standard 7-stage task (Chase et al., 2012; Tait et al., 2014).

It does not seem to be important that the apparatus and materials are standardised for the rat ID/ED ASST: research groups typically construct their own testing chamber or arena, and the various elements of the stimuli (odours and digging media) are largely determined by local availability. On the one hand, this variability indicates the robustness of the task, nevertheless there are aspects that are important to consider when selecting materials. Here we will, therefore, discuss some of the reasoning behind choices made during the development of the rat ID/ED task, including changes made since the original Birrell and Brown publication so that researchers wishing to adopt or adapt the task in the future are informed by our experience of what worked or did not work. We will focus on designs for use with rats, as mouse ASSTs, although often similar in design to the rat ASST, have their own requirements (see Tait et al., 2014 for review). In this methods paper, we will discuss rat strain; the apparatus; the choice of stimulus exemplars; stage and trial order; counterbalancing; and data analysis.

Section snippets

Rats

ASST data have been collected in many different rat strains – including Lister Hooded (Birrell and Brown, 2000); Long Evans (Rodefer et al., 2008); Sprague Dawley (Tunbridge et al., 2004); Wistar Kyoto (Cao et al., 2012); Fischer 344/Brown Norway cross (McCoy et al., 2007) – and although the majority of the published work has used male rats, there are also data from female rats (Lovic and Fleming, 2004; McLean et al., 2012). Whereas there may be some strain or sex differences in willingness to

The testing arena

The ASST arena used by Birrell and Brown (2000), and in all our subsequent studies, is made from a modified opaque white plastic homecage, and has overall outer dimensions of 70 × 40 × 18 cm (Fig. 1). It is divided into three sections: roughly two thirds of the length of the arena serves as the holding area, separated from the remaining third by removable semi-transparent acrylic barriers, which is further subdivided by an opaque white plastic wall, creating two choice chambers. Access to

Training

As foraging is a spontaneous behaviour, training is not a requirement. However, familiarising the rats with digging in the bowls for food does speed subsequent testing. Furthermore, rats show food neophobia, and therefore familiarisation with the cereal, which is a novel food, is necessary (Modlinska et al., 2015). Typically, rats are given a small quantity of the cereal to be used as reward the day before they are exposed to the reward-baited bowls. Although this form of habituation is not

Conclusion

When well-implemented, the rat ASST provides a robust and repeatable measure of the cognitive processes involved in attentional set-shifting behaviour and thus convergent validity with the CANTAB ID/ED task for humans and monkeys. Whilst throughput can initially be limited by time taken to conduct the task, with repeated testing and within-subjects designs, the task can be completed in under an hour, allowing dose response curves to be established for acute pharmacological interventions, as

Funding

The initial development of the attentional set-shifting task was supported by The Wellcome Trust (Project Grant 051945/Z/97/Z) and a Biotechnology and Biological Sciences Research Council (UK) Studentship to Jennifer M. Birrell. The production of this manuscript was supported in part by a Faculty Development Grant awarded to LSN.

Conflicts of interest

None.

Acknowledgements

The authors would like to acknowledge the invaluable assistance of Drs Jennifer Birrell, Andrew Blackwell, Kerry McAlonan, Shuang Xia, Alonzo Whyte, Ana Garcia Aguirre, Rudi Stanislaus-Carter, as well as PhD students Sandeep Dhawan and Jiachao Wang, and numerous undergraduate students from the University of St Andrews, for the many hours spent collecting and analysing the behavioural data that contributed to our database of control rats.

References (79)

  • M.D. Lapiz et al.

    Noradrenergic modulation of cognitive function in rat medial prefrontal cortex as measured by attentional set shifting capability

    Neuroscience

    (2006)
  • H.S. Lindgren et al.

    Lesions of the dorsomedial striatum impair formation of attentional set in rats

    Neuropharmacology

    (2013)
  • V. Lovic et al.

    Artificially-reared female rats show reduced prepulse inhibition and deficits in the attentional set shifting task--reversal of effects with maternal-like licking stimulation

    Behav. Brain Res.

    (2004)
  • K. McAlonan et al.

    Orbital prefrontal cortex mediates reversal learning and not attentional set shifting in the rat

    Behav. Brain Res.

    (2003)
  • J. McGaughy et al.

    Noradrenergic, but not cholinergic, deafferentation of prefrontal cortex impairs attentional set-shifting

    Neuroscience

    (2008)
  • S.L. McLean et al.

    A preliminary investigation into the effects of antipsychotics on sub-chronic phencyclidine-induced deficits in attentional set-shifting in female rats

    Behav. Brain Res.

    (2008)
  • K. Modlinska et al.

    Food neophobia in wild and laboratory rats (multi-strain comparison)

    Behav. Processes

    (2015)
  • W.M. Mohamed et al.

    Methylphenidate improves cognitive deficits produced by infantile iron deficiency in rats

    Behav. Brain Res.

    (2011)
  • K.L. Murphy et al.

    Exposure to sevoflurane anesthesia during development does not impair aspects of attention during adulthood in rats

    Neurotoxicol. Teratol.

    (2017)
  • A. Nikiforuk et al.

    Mazindol attenuates ketamine-induced cognitive deficit in the attentional set shifting task in rats

    Eur. Neuropsychopharmacol.

    (2010)
  • A.M. Owen et al.

    Extra-dimensional versus intra-dimensional set shifting performance following frontal lobe excisions, temporal lobe excisions or amygdalo-hippocampectomy in man

    Neuropsychologia

    (1991)
  • E.M. Powell et al.

    Cognitive flexibility: development, disease and treatment

    Neuroscience

    (2017)
  • D. Scheggia et al.

    The ultimate intra-/extra-dimensional attentional set-shifting task for mice

    Biol. Psychiatry

    (2014)
  • D.S. Tait et al.

    Lesions of the basal forebrain impair reversal learning but not shifting of attentional set in rats

    Behav. Brain Res.

    (2008)
  • D.S. Tait et al.

    Tacrine improves reversal learning in older rats

    Neuropharmacology

    (2013)
  • D.S. Tait et al.

    Effects of lesions of the subthalamic nucleus/zona incerta area and dorsomedial striatum on attentional set-shifting in the rat

    Neuroscience

    (2017)
  • M.D. Barense et al.

    Aged rats are impaired on an attentional set-shifting task sensitive to medial frontal cortex damage in young rats

    Learn. Mem.

    (2002)
  • M.G. Baxter

    Age-related effects on prefrontal cortical systems: translating between rodents, nonhuman primates, and humans

  • E.A. Berg

    A simple objective technique for measuring flexibility in thinking

    J. Gen. Psychol.

    (1948)
  • J.M. Birrell et al.

    Medial frontal cortex mediates perceptual attentional set shifting in the rat

    J. Neurosci.

    (2000)
  • G.B. Bissonette et al.

    Double dissociation of the effects of medial and orbital prefrontal cortical lesions on attentional and affective shifts in mice

    J. Neurosci.

    (2008)
  • J.M. Brooks et al.

    Transient inactivation of the neonatal ventral hippocampus impairs attentional set-shifting behavior: reversal with an alpha7 nicotinic agonist

    Neuropsychopharmacology

    (2012)
  • V.J. Brown et al.

    Behavioral flexibility: attentional shifting, rule switching, and response reversal

    Encyclopedia Psychopharmacology

    (2015)
  • V.J. Brown et al.

    Attentional set-shifting across species

    Curr. Top. Behav. Neurosci.

    (2016)
  • A.H. Cao et al.

    Effects of methylphenidate on attentional set-shifting in a genetic model of attention-deficit/hyperactivity disorder

    Behav. Brain Funct.

    (2012)
  • E.A. Chase et al.

    Lesions of the orbital prefrontal cortex impair the formation of attentional set in rats

    Eur. J. Neurosci.

    (2012)
  • H.F. Clarke et al.

    Prefrontal serotonin depletion affects reversal learning but not attentional set shifting

    J. Neurosci.

    (2005)
  • R. Dias et al.

    Primate analogue of the Wisconsin card sorting test: effects of excitotoxic lesions of the prefrontal cortex in the marmoset

    Behav. Neurosci.

    (1996)
  • P.A. Dudchenko et al.

    Neurotoxic hippocampal lesions have no effect on odor span and little effect on odor recognition memory but produce significant impairments on spatial span, recognition, and alternation

    J. Neurosci.

    (2000)
  • Cited by (47)

    • Adolescent female valproic acid rats have impaired extra-dimensional shifts of attention and enlarged anterior cingulate cortices

      2023, Brain Research
      Citation Excerpt :

      For this experiment multiple ID phases were used to induce the ED effect (Tait et al., 2018). The ASST relies on a wide neural network of brain regions for correct performance, with medial prefrontal regions mediating ED shifts (Tait et al., 2018; Birrell and Brown, 2000) and the anterior cingulate mediating ID shifts (Bubb et al., 2020; Ng et al., 2007). A recent publication demonstrates that VPA rats were impaired when performing the set-shifting task, which was designed to capture executive function deficits observed in humans (Birrell and Brown, 2000).

    • Taurine in symptom amelioration and recovery in lead-induced neurotoxicity

      2023, Treatments, Nutraceuticals, Supplements, and Herbal Medicine in Neurological Disorders
    View all citing articles on Scopus
    View full text