Mental Imagery in the Science and Practice of Cognitive Behaviour Therapy: Past, Present, and Future Perspectives

Mental imagery has a long history in the science and practice of cognitive behaviour therapy (CBT), stemming from both behavioural and cognitive traditions. The past decade or so has seen a marked increase in both scientific and clinical interest in mental imagery, from basic questions about the processes underpinning mental imagery and its roles in everyday healthy functioning, to clinical questions about how dysfunctions in mental imagery can cause distress and impairment, and how mental imagery can be used within CBT to effect therapeutic change. This article reflects on the current state of mental imagery in the science and practice of CBT, in the context of past developments and with a view to future challenges and opportunities. An ongoing interplay between the various strands of imagery research and the many clinical innovations in this area is recommended in order to realise the full therapeutic potential of mental imagery in CBT.

Mental imagery has played a role in cognitive behaviour therapy (CBT) throughout its history, but the past decade or so has seen an acceleration of interest in the use of mental imagery across an ever-growing range of disorders and treatment applications (Saulsman et al. 2019). Several converging factors have likely contributed to this rising interest in and use of imagery. First, there has been a broader scientific resurgence of interest in mental imagery, facilitated both by the advent of better neuroimaging techniques, allowing insights into the neural underpinnings of imagery (Kosslyn et al. 2001;Pearson et al. 2015), and by experimental work supporting the perceptual nature of much emotional memory (e.g. Andrade et al. 1997;Arntz et al. 2005;Brewin and Saunders 2001), and testing the assumption of a special relationship between imagery and emotion (e.g. Holmes and Mathews 2005). Second, there has been growing recognition and documentation of the presence of mental imagery dysfunctions across the whole range of psychopathology (e.g. Holmes and Hackmann 2004). Third, there has been increasing clinical interest in the potential applicability of methods developed within the context of trauma memories, such as imagery rescripting, to other kinds of distressing memories and problematic imagery more broadly (e.g. Arntz and Weertman 1999;Holmes et al. 2007a). This convergence of scientific and clinical interest has led to a particularly fertile interplay between basic research and applied clinical work.
This article reflects on the current status of mental imagery in the science and clinical practice of CBT, in the context of past developments, and with a view to future challenges and opportunities. After a brief introduction to mental imagery, this article will consider past and present developments of relevance to CBT, organised into the following four broad categories of work spanning science and practice: the use of mental imagery as an experimental tool to test and develop underlying theories; understanding the role of mental imagery in healthy everyday functioning; understanding of the role of mental imagery dysfunctions in psychopathology; and the use of mental imagery as a therapeutic tool within CBT itself. Finally, building on this overview of the present state of the field, the article will consider future perspectives for imagery in CBT research and treatment development. The article is not intended as a comprehensive review (and does not consider broader applications of imagery within the CBT context, for example to therapist self-care, training, or supervision e.g. Bennett-Levy et al. 2009;Prasko et al. 2020), but rather highlights selected examples of past and current work to illustrate broader patterns and trends in the field. The 'present status' of CBT-relevant mental imagery research is very much informed by the programme of the 2019 World Congress of Behavioural and Cognitive Therapies (WCBCT; , at which the numerous presentations on the topic of mental imagery spanned disorder areas including depression, anxiety disorders, addictions, and psychosis, illustrating the breadth of clinical interest at the present time. By providing an overview of where mental imagery stands at the current crossroads of science and clinical practice, the article aims to stimulate reflection and promote continued interchange across the many strands of work relevant to understanding how we can make best use of mental imagery in CBT.

Why Focus on Mental Imagery?
Mental imagery can be defined as "representations and the accompanying experience of sensory information without a direct external stimulus" (Pearson et al. 2015, p. 590), or in more colloquial terms as "'seeing with the mind's eye,' 'hearing with the mind's ear,' and so on" (Kosslyn et al. 2001, p. 635). Most of us will be familiar with the experience of mental imagery in our daily lives, for example when we recall an event from memory and replay the scene in our mind, or when we think forward to an upcoming event and 'pre-experience' some anticipatory emotion, such as excitement or dread, as we play out in our mind's eye the different ways in which the event could potentially unfold. Many of us will also be familiar with the problematic imagery reported by patients or clients in clinical practice, for example distressing memories that return uncontrollably again and again, or upsetting scenes of anticipated catastrophes or disappointments. However, despite the ubiquity of mental imagery, clinicians, researchers, and patients may not always be aware of the scientific basis for considering it as a form of thought deserving particular attention in CBT (e.g. Bell et al. 2015;Blackwell 2019).
From a basic research perspective, experimental investigations have shown that imagery-based thought can have particularly strong effects on emotion, cognition, and behaviour compared to non-imagery (e.g. verbal) thought (see e.g. Blackwell 2019; Holmes and Mathews 2010, for overviews). This is most likely due to the nature of the representation of imagery at a neural level, which is very similar to that for actual perception of sensory input (Pearson et al. 2015), and leads to imagery having an 'as-if reality' quality (Ji et al. 2016). From a more clinical perspective, the presence of dysfunctional imagery has been documented across an increasing range of disorders , and thus neglecting imagery in theoretical conceptualisations or clinical assessments risks excluding a potential key contributor to distress, functional impairment, and symptom maintenance. This point is particularly important given that patients may not report imagery unless it is explicitly asked about (e.g. Beck et al. 1979;Hales et al. 2014). Finally, an awareness of mental imagery, its functions, and properties, open up opportunities for a range of powerful therapeutic applications, from incorporation of imagery into 'standard' therapy tools such as thought records (Josefowitz 2017), to use of imagery-based techniques such as imagery rescripting (e.g. Arntz 2012; Strachan et al. 2020), and even complete CBT treatment programmes that have a specific focus on imagery (e.g. Holmes et al. 2019;Jung and Steil 2013;McEvoy and Saulsman 2014;Taylor et al. 2019). Arguably, discussion of almost any psychological process, psychopathological phenomenon, or treatment technique is incomplete unless the potential role of mental imagery has been explicitly considered (Blackwell 2020).

Mental Imagery Across the Science and Practice of CBT
The next sections will consider four ways in which mental imagery is of relevance for the science and practice of CBT: use of mental imagery as an experimental tool; understanding the role of mental imagery in healthy everyday functioning; mental imagery dysfunctions in psychopathology; and the clinical application of mental imagery-based techniques within the practice of CBT.

Mental Imagery as an Experimental Tool
The practice and continued development of CBT is rooted in theories and models of the role of cognition, behaviour, emotion, and physiology in the maintenance of psychopathology, and these theories and models can be tested and informed by scientific investigation. From a basic science perspective, the properties of mental imagery, such as its effect on emotion, cognition, and behaviour (Blackwell 2019), and more broadly, ability to serve as a simulation of reality (Ji et al. 2016), mean that imagery can provide a valuable experimental tool to probe numerous aspects of the theories underpinning CBT approaches.
One example of where the ability of mental imagery to evoke emotional and physiological responses has long been capitalised upon is in the context of fear research, for example using imagery scripts as a lab-based analogue of actual exposure to a feared situation (e.g. Lang et al. 1980). In fact, many of these lab-based applications of imagery followed on from clinical observations of the surprising effectiveness of imaginal exposure in treating anxiety disorders, which led to the development of Peter Lang's bio-informational theory of imagery (Ji et al. 2016;Lang 1977Lang , 1979. Thus, imagery scripts could be used to explore differential physiological responses to different kinds of situations, such as phobia-related, neutral, or positive scenarios, amongst people with different kinds of anxiety disorders in a way that might not be possible 'in vivo'. Coupled with interest at the time in physiological responses to imagery in the context of treatment for anxiety disorders, and potential differences between imagery-based versus in vivo exposure (e.g. Marks et al. 1971;Mathews 1971), the controlled use of imagery scripts in laboratory situations represented a close connection between science and clinical practice. Interestingly, such research reflects an experimental use of imagery to understand aspects of disorders that are not imageryrelated in themselves (e.g. fear responses). Imagery can of course also be used as an experimental tool to test theoretical statements about the role of imagery itself in disorders, for example the effects of negative observer-perspective imagery in the context of social phobia (e.g. Hirsch et al. 2005;Spurr and Stopa 2003), the potential role of imagery in driving craving in the context of substance use disorders (e.g. Harvey et al. 2005;May et al. 2004), or as an analogue of trauma memories or other distressing imagery to test the effects or mechanisms of potential therapeutic techniques (e.g. Engelhard et al. 2011).
While there is a multitude of ongoing work using imagery to test and build theory in experimental contexts, one area with particular relevance for CBT that has seen a recent re-ignition of interest is the use of imagery in fear conditioning paradigms (see Mertens et al. 2020). Fear conditioning, extinction, and related paradigms have been widely used to test and refine theories of anxiety disorders, including their development, remission, and treatment. Careful use of mental imagery within these paradigms offers a number of opportunities to increase their ecological validity and relevance to psychopathology and its treatment . For example, research has shown that fear responses to stimuli, including avoidance , can be conditioned when aversive contingencies are only ever imagined and not experienced in reality (Mueller et al. 2019;Soeter and Kindt 2012). Mental images themselves can act as a conditioned stimulus, with conditioned responses generalising to actual stimulus presentation (Lewis et al. 2013). Further, imagery-based recall can provide a means to reactivate fear memories, potentially rendering them vulnerable to disruption (Grégoire and Greening 2019). An interesting parallel comes from work investigating mechanisms underlying the development of intrusive memories of trauma. This research has long used film stimuli as an analogue of a traumatic event in order to induce involuntary memories and study the variables modulating their characteristics and occurrence (James et al. 2016), but in fact, intrusive memories can also be induced when the experimental 'trauma' is never witnessed (via film) but only imagined (e.g. Krans et al. 2009;Mooren et al. 2019). The relevance to CBT of using mental imagery in conditioning and related paradigms becomes clear if we consider how much of our emotional world is essentially internal and rich in mental imagery, how many disorders are characterised by fears of events that did not or have not yet happened, but have only been 'experienced' via imagery, and how much imaginal rehearsal of negative contingencies occurs in the context of processes such as rumination and worry. Thus, these lines of research hold exciting potential for future developments in our theoretical understanding of anxiety disorders, as well as having clear links to potential clinical applications.

Mental Imagery in Healthy Everyday Functioning
A second way in which the science of mental imagery can inform CBT is via research uncovering the role of mental imagery in everyday healthy functioning. That is, the more we know about when we experience imagery in daily life, and the functions such imagery serves, the better able we may be to identify dysfunctions in such imagery or deficits that can be addressed within therapy. The study of 'normal' mental imagery in fact goes back to the beginnings of experimental psychology itself (e.g. Fechner 1860; see Roeckelein 2004 for an overview), and although investigations of mental imagery have gone hand in hand with discussions of its functions, for example in memory, problem-solving, or broader aspects of conscious experience (e.g. Betts 1909;Galton 1883;Marks 1999), this has not necessarily fed into CBT theory or practice (apart from, perhaps, where it has informed deliberate use of imagery within some therapeutic techniques e.g. Meichenbaum 1978). One exception might be the role of specific memory retrieval in planning and problem-solving (Williams 2006), whereby deficits in this are exhibited in overgeneral memory in depression (e.g. Sumner et al. 2010;Williams and Broadbent 1986) and are linked to various dysfunctions, such as in interpersonal problem-solving (e.g. Raes et al. 2005). This has informed approaches such as memory specificity training (Raes et al. 2009), and overgeneral memory has also been a target for mindfulness-based cognitive therapy (Williams et al. 2000). Although not conceived as a mental imagery dysfunction per se, there is an intrinsic link between overgeneral memory and mental imagery due to the role of imagery in retrieval processes and the experience of episodic memory recall (see e.g. Holmes et al. 2016a, for a discussion).
However, the past decade or so has seen an ever-increasing amount of research shedding light on the role of mental imagery in everyday functioning, in particular in relation to memory and future-oriented thinking, in a way that has clear implications for CBT. Contributions have come both from a neural mechanisms perspective (e.g. Addis and Schacter 2008;Hassabis and Maguire 2009;Miloyan et al. 2019;Schacter et al. 2012), and from studies inducing or recording the experience of voluntary and involuntary memories and futureoriented thoughts in controlled laboratory settings or daily life (e.g. Barzykowski and Niedźwieńska 2018;Berntsen 1996;Berntsen and Jacobsen 2008;Cole and Kvavilashvili 2019;D'Argembeau et al. 2011;Mace 2005). Although this research is often not framed from a mental imagery perspective, the memories, future projections, and other spontaneous thoughts recorded in such studies often have a rich imagery component. Further, these imagery-rich thoughts, whether deliberately generated or spontaneously occurring, appear to be common in daily life and are implicated in a number of important roles, for example in planning, decision making, motivating or prompting behaviour, maintaining one's selfimage, and emotion regulation. To illustrate this more concretely, if someone receives a message from a friend inviting them to a party the upcoming weekend, they might find themselves automatically imagining being there and chatting or dancing with their friends; how enjoyable this 'feels' in their imagination may influence whether they accept the invitation or find an excuse not to make it (imagery in decision making). If this image of being at the party generates particularly strong positive emotions, the individual may deliberately bring it to mind throughout the week to improve their mood when they are struggling at work (imagery in emotion regulation). Walking past a clothes shop the image may pop spontaneously to mind and prompt the individual to enter the shop and buy a new outfit for the party (imagery prompting behaviour). When the evening of the party comes, perhaps it is freezing cold and pouring with rain, such that the individual feels like cancelling and staying at home; however, imagining how much fun the party will be and 'pre-experiencing' this enjoyment (or perhaps the friend's disappointment and subsequent complaining if they do not come) might shift the balance in favour of going out (imagery motivating behaviour). Before leaving the house, the individual may mentally go through their usual route to the party location in their mind, but 'seeing' the normally pleasant path through the park in the dark makes them consider taking a different route (imagery in planning). Finally, arriving at the party and seeing an old friend they have not met in years may trigger re-experiencing of memories of past shared experiences, activating a sense of being a loyal and valued friend (imagery in self-image). Of course, all of this may occur without experiencing much or any imagery at all. However, emotional thoughts often have some imagery-based component (Moritz et al. 2014), and engaging in imagery-rich thinking, such as 'mental time travel' into the past and future, appears to be almost a 'default' activity that our minds engage in when we are not occupied in task-directed thought (e.g. Schacter et al. 2007). Hence, biases, deficits, or disruptions in the frequency, quality, or content of such imagery could have far-reaching consequences and readily contribute to psychopathology, and provide suitable targets for treatments (Blackwell 2019(Blackwell , 2020. The research cited above generally investigates the potential roles of imagery in healthy functioning via examining the circumstances and effects of imagery's presence. An interesting complement to this comes from research examining the effect of imagery's absence. The effect of not thinking in images, but predominantly in verbal-linguistic forms, has long been studied in relation to worry and generalised anxiety disorder (e.g. Borkovec et al. 1993). However, recently there has been increased research interest in the phenomenon of aphantasia, in which individuals have little or no conscious experience of mental imagery (e.g. Dawes et al. 2020;Zeman et al. 2015). Research shows, for example, that individuals with aphantasia have different patterns of emotional response to written text (Wicken et al. 2019), highlighting a crucial role for mental imagery in generating emotional responses to emotional (text-based) information. Developing our understanding of how mental imagery occurs and its functions in everyday healthy functioning-whether by studying imagery's presence or its absence-can help not only inform theories of dysfunction, but also development of intervention strategies to help compensate for or rebuild potential deficits in such functions where they appear to contribute to clinical impairment or distress.

Dysfunctional Mental Imagery
A third way in which the science of mental imagery can inform CBT is via the discovery, documentation, and exploration of dysfunctional experiences of mental imagery in psychopathology. This strand has perhaps the longest history, in that descriptions of people experiencing intrusive memories or imagined catastrophes are found throughout both historical clinical reports and literature more broadly. In fact, early cognitive therapy manuals explicitly noted the importance of enquiring about the occurrence of images (e.g. Beck 1976), and dysfunctional images have long been included in cognitive models not only of disorders in whose phenomenology they obviously play a key role, such as PTSD and obsessive-compulsive disorder, but also others where their role is more subtle, such as social phobia (e.g. Clark and Wells 1995;Rapee and Heimberg 1997). However, it is only relatively recently that detailed and systematic documentation and investigation of dysfunctional mental imagery across a wider range of disorders has started to occur (Holmes and Hackmann 2004).
Subsequently, over the past decade or so, it has become apparent that dysfunctions in mental imagery are widespread across psychopathology, from the experience of intrusive memories (e.g. Reynolds and Brewin 1999;Williams and Moulds 2008) or suicidal 'flashforwards' in the context of depression (e.g. Holmes et al. 2007b), to a broad spectrum of imagery dysfunctions across a wide range of disorders, including psychosis (Malcolm et al. 2015), bipolar disorder (Di Simplicio et al. 2016;Hales et al. 2011), worry and generalised anxiety disorder (Hirsch et al. 2006;Tallon et al. 2020), eating disorders (Dugué et al. 2016;Kadriu et al. 2019), and incontinence phobia (Pajak et al. 2013)-essentially, in the context of any disorder or clinical manifestation of distress in which this has been investigated. Importantly, although the presence of distressing images may be the most obvious example of mental imagery dysfunction, problems can also occur in the form of the absence or reduced quality of beneficial or adaptive imagery, for example difficulties in generating positive future imagery as observed in depression (Holmes et al. 2016a), which have been linked to core features of the disorder such as reduced anticipatory pleasure (Hallford et al. 2020a).
Of course, it is important not only to document the presence (or absence) of imagery in the context of psychopathology, but also to fully characterise such imagery and its consequences. This has been an increasing focus of recent research, for example via obtaining finer-grained descriptions of different facets of dysfunctional imagery (e.g. non-visual, such as olfactory or somatosensory, sensory modalities; Dobinson et al. 2020;Weßlau et al. 2016) and delineating its roles more precisely, for example linking the occurrence of intrusive imagery more closely to core aspects of functioning such as the self (Çili and Stopa 2015). Recent developments in this area have also included increased interest in the potential role of imagery dysfunctions amongst children and adolescents (e.g. Chapman et al. 2020;Pile and Lau 2020), other adult populations such as those with brain injury , and explicit incorporation of imagery into a wider range of psychological models of disorders and disordered behaviour, such as suicide (O'Connor and Kirtley 2018). Alongside research into the role and functions of mental imagery in everyday healthy functioning, as mentioned in the previous section, this growing awareness of imagery dysfunctions across psychopathology can contribute to a comprehensive picture of imagery from health to disorder. Further, once mental imagery dysfunctions have been identified in a particular disorder or clinical presentation, these can then be targeted in therapy, as will be discussed in the next section.

Mental Imagery as a Therapeutic Tool within CBT
Mental imagery can be used in a multitude of ways within CBT, including as an assessment tool (e.g. imaginal reliving of a situation to identify automatic thoughts), within imagery-focussed techniques such as imaginal exposure or imagery rescripting, or as an additional component within techniques for which imagery is not itself the main focus, such as problem-solving or challenging negative automatic thoughts. There are several historical overviews of the development of imagery techniques in CBT and psychological therapy more broadly (Edwards 2007(Edwards , 2011Singer 2006). What is interesting to note from such histories is the ubiquity of imagery not only across psychotherapies in general, but also specifically in the early development of cognitive and behavioural therapies. These included central roles of imagery in systematic desensitisation (Wolpe 1958), forms of rehearsal such as 'stress inoculation' (Meichenbaum 1974), 'covert modelling' (Kazdin 1973), or 'rational restructuring' (Goldfried and Davison 1976), and various aspects of the treatment of anxiety disorders and depression within cognitive therapy (Beck et al. 1979;Beck and Emery 1985). The development of mental imagery within CBT to some extent mirrors the development of CBT more generally, in that one major influence comes from the behavioural tradition, for example via imaginal exposure (Wolpe 1958), and another out of a more cognitive tradition such as the work of Beck (e.g. Beck et al. 1979). These traditions come together in later work where the focus is explicitly on behavioural experiments-active behaviour-based testing out of cognitions or beliefs (Bennett-Levy et al. 2004)-for example in using video-recording of a speech task to test negative beliefs about the accuracy of a self-image in social phobia (e.g. McManus et al. 2009;Warnock-Parkes et al. 2017).
However, the claim of a resurgence in interest in imagery within CBT in recent years implies that there was a dip in such interest at some point; certainly, reading through CBT manuals from the past 20-30 years, one would often get the impression that mental imagery was not figuring largely in clinicians' or researchers' thinking. It may be that there was a relative focus on other areas or that imagery simply became 'crowded out' via development of and preference for other techniques. Whatever the reason, it is definitely the case that imagery-based techniques can now be found across an increasingly broad spectrum of CBT applications. One example comes through the use of imagery rescripting across an ever-growing range of disorders and problem areas (Arntz 2012). Imagery rescripting most commonly involves reliving a distressing memory but imagining events turning out differently-for example, fighting back or being rescued from an attacker in case of an assault, or viewing the event through the eyes of an adult and intervening in the case of childhood abuse (e.g. Holmes et al. 2007a; Raabe et al. 2015). Within the CBT tradition, imagery rescripting was developed in the context of memories of childhood trauma, and schema therapy for people with diagnoses of personality disorders; the rationale, or at least one hypothesised mechanism of action, was that via incorporation of corrective information via imagery, the meanings, appraisals, and emotions associated with the distressing memory would be modified. Although originally applied to distressing or core dysfunctional memories, in fact, the technique can be applied to any kind of imagery, and the past decade has seen an ever-increasing range of its applications as clinicians and researchers have realised the therapeutic possibilities. This has included using imagery rescripting of distressing memories as a stand-alone treatment for depression (Brewin et al. 2009 Alongside the increased application and development of specific techniques, such as imagery rescripting, recent years have also seen greater attention paid to imagery within CBT approaches more generally (e.g. increased consideration of imagery within the second edition of Mind over Mood; Greenberger and Padesky 2015). Further, a range of CBT therapies have been developed where imagery is the main or even sole focus, for example in imagery-focused CBT approaches for social anxiety (McEvoy and Saulsman  These imagery-focussed treatments conceptualise specific experiences of imagery as being central to a core symptom or component of a disorder; broadly speaking, it then follows that a targeted focus on this imagery-based mechanism holds potential to bring about substantial improvements. In addition to the expansion and refinement of imagery-focussed CBT approaches that target maladaptive imagery, another recent trend has been in the increased development of various techniques designed to foster positive or adaptive imagery. Some examples come from the context of depression, in which a difficulty imagining positive future events has been linked to symptoms such as anhedonia and lack of motivation. Approaches have included practice in generating vivid and specific imagery of future events (Hallford et al. 2020b;Hallford et al. 2020c) or rewarding experiences (Linke and Wessa 2017), or computerised cognitive training approaches involving repeated generation of positive imagery (e.g. Blackwell et al. 2015;Dainer-Best et al. 2018). Other examples not restricted to depression build on research indicating a role for imagery in goal-directed behaviour. These aim to increase engagement in healthy desired behaviour, for example via imagery of the steps required to carry out the behaviour and of the rewarding outcome (e.g. Renner et al. 2019;Solbrig et al. 2018). Further, there is continued interest in examining the effects and potential benefits of compassion-focussed imagery across a range of applications (e.g. Campbell et al. 2019;McEwan and Gilbert 2016;Naismith et al. 2019). In fact, there are now a huge variety of approaches involving repeated rehearsal of positive imagery or memories that are in various stages of development and testing (e.g. see Hitchcock et al. 2017, for a review and meta-analysis including a range of these approaches).
The sheer amount of work exploring different ways of using imagery in therapy is testament to the promise seen in imagery-focussed techniques and approaches. However, apart from a small number of cases in which imagery-based work is part of an evidence-based treatment (e.g. prolonged exposure or trauma-focussed CBT for PTSD; Ehlers et al. 2005;Foa et al. 2007), most of these imagery-based treatments or techniques are still under development and evaluation; which of these should ultimately be integrated into routine clinical practice, and how best this can be achieved, remain open questions for future research.

Future Perspectives
As outlined in this paper, the past decade or so has seen an explosion of interest in mental imagery across the whole CBT spectrum, from basic research to treatment innovation; this is manifested in an ever-increasing proliferation of new research papers and opens up many exciting future possibilities. Of course, within each specific line of research-whether using imagery to probe theory, investigating functions of healthy mental imagery, uncovering imagery dysfunctions, or developing new treatment techniques-there will be many areas of particular promise and future directions that could be highlighted. Further, across all of these lines of research and clinical innovation, there is a great need for much more work focussing on developmental, life-span, and cross-cultural aspects of imagery. However, in line with the aim of this paper to provide an overview of the field, this final section will take a step back from the details of the work discussed in the previous sections and instead consider a broader perspective on the future of imagery in the science and practice of CBT. This will start by addressing research perspectives and then move on to clinical innovation and implementation, before considering what this means for the individual researcher or clinician.

Research Perspectives
There is now a vast breadth of research examining mental imagery from neural to behavioural levels, and ideally the development and optimisation of imagery-based treatment techniques in CBT would draw on this rich source of information to inspire new, and hone existing, approaches. However, the sheer quantity and variety of the research can be overwhelming, and within the literature, there is often little true integration between different levels of understanding or perspectives on imagery (e.g. through psychopathology, healthy functioning, and neural levels of explanation) beyond passing references in the introduction and discussion sections of papers. Further, because mental imagery itself comes into play across so many psychological processes, much research of relevance for understanding imagery and its functions is spread across work in memory, planning, goals, decision making, and other aspects of cognition in papers that may often make no explicit reference to the imagery-related components of cognition involved in these functions. There are also many overlaps between imageryfocussed research and other concepts such as embodied cognition (e.g. Palmiero et al. 2019), and with other non-clinical fields of psychology such as sports psychology (e.g. Cumming and Williams 2012); these fields may use different or only partially overlapping terminologies to describe similar phenomenon, and may sometimes even represent repetitions of the same research in parallel under different names. Such diversity in concepts and terminology is not itself necessarily a problem; in fact, it can provide additional perspectives leading to new insights and ideas, as well as conceptual replications of important phenomena. However, it provides a challenge to researchers or clinicians trying to gain a comprehensive understanding of the science underpinning mental imagery and how this might inform clinical practice.
Not only is the mental imagery literature characterised by this vast breadth of research, but there has also been an ever-increasing focus on finer and finer details within each specific field of investigation. This attention to detail is necessary and highly important. For example when enquiring about the presence of imagery or asking someone to imagine a particular scene, the precise nature of the imagery being experienced or to be generated needs to be closely specified; research needs to avoid using imagery as a blunt tool (simply asking participants to imagine something without specific instructions as to how) but rather as a scalpel or probe aiming for pinpoint-precision to answer the more specific questions that are increasingly of interest. From a basic research perspective, a finer-grained understanding of imagery and its impact can benefit from methodological advances, such as the increasing accessibility and convenience of ecological momentary assessment (EMA) approaches to capture spontaneous thoughts and images in (close to) real-time (e.g. Beaty et al. 2019;Slofstra et al. 2017), or use of virtual reality to create more ecologically valid but highly controlled environments for the encoding (e.g. Schweizer et al. 2018) and retrieval (e.g. Zlomuzica et al. 2018) of image-based memories. However, given the essentially internal and subjective nature of the experience of mental imagery, precision and attention to detail in constructing task instructions, or in eliciting descriptions of images from participants or patients, will always be a pre-requisite for informative research in mental imagery, and can in itself go a long way to achieving precisely-tailored imagery experiences in experimental settings and clinical settings. Similarly, within a treatment context, there undoubtedly remain many important insights that can be gained through careful clinical observation and questioning, which may then feed into new research questions or treatment techniques.
How can the needs for both inter-disciplinary integration and an ever-increasing focus on fine detail, as discussed above, be reconciled? A usable mental imageryfocussed 'grand theory of everything' aiming to integrate all our research knowledge is probably not only unfeasible but also not necessarily useful for many practical purposes. One more manageable alternative is to focus on specific, restricted, phenomena thought to have key clinical relevance, and within this specific focus try to incorporate and synthesise the relevant research findings and insights from other fields, disciplines, and perspectives, thus drawing on the richest possible information to understand and treat the specific phenomenon under investigation. The experience of distressing intrusive imagery is one example where this kind of approach has been explicitly discussed (e.g. Singh et al. 2020;Visser et al. 2018). If an ultimate aim of the research is to improve treatment outcomes, then applying a comprehensive inter-disciplinary perspective within a narrow focus on a specific clinically relevant imagery target may provide a feasible way to synthesise many sources of information and perspectives without losing sight of important details.

Clinical Innovation and Implementation Perspectives
Clinical innovation in the use of imagery in CBT may encounter a similar challenge to that facing research in this area: There is now simply so much published that it can be difficult to get a good grasp of what exactly has been done before and why it might have been effective or otherwise. Given the long-standing use of imagery in CBT and psychological therapy more broadly, 'new' techniques or approaches may sometimes be presented that appear to be very similar to existing but perhaps neglected ideas from the past, leading to a sense of re-invention of the wheel. This is not a phenomenon that is limited to imagery, but is well-illustrated in this area given the use of imagery across so many different formats of psychotherapy and, superficially at least, the limited scope of things one might ask someone to imagine. Thus, someone might read about a new technique and think: isn't this just imaginal exposure under another name, or with some extra 'features' that may or may not add anything? Although wheels are probably reinvented at regular intervals within psychological therapy research and practice, particularly given that certain techniques or ideas tend to come in and out of fashion over time, it is also important to consider that a certain amount of re-appraisal of 'old' techniques is inevitable as a field develops, and is a necessary first step before improving such techniques: theories and models develop over time (hopefully), and before a technique can be applied and improved it needs to be explainable within the framework being used by the researcher or clinician. As an extreme example, a clinician or researcher may discover a technique used by Freud that, when they try it, is amazingly effective with their patients. However, they are unlikely to take Freud's rationale for the treatment's use but rather will need to re-explain it within their own working therapy model (perhaps adjusting the model if required). This in turn will likely suggest different ways to improve its application than those that would have been suggested by Freud's initial conceptualisation of the treatment technique's mechanism. As a perhaps more subtle example from within imagery work, Wolpe conceptualised the role of relaxation in his systematic desensitisation procedure as providing 'reciprocal inhibition' (Wolpe 1958). However, a later suggestion was that if relaxation had a beneficial effect in imaginal exposure, this may have been via allowing patients to generate more vivid imagery (Mathews 1971); this alternative conceptualisation would lead to quite different approaches to improving the treatment's effectiveness. To take a more recent example unrelated to imagery, viewing exposure through the lens of inhibitory learning rather than previous ideas around habituation (Craske et al. 2014) also leads to divergent routes for improving this technique, albeit from the same procedural starting point. Thus some degree of re-invention of older techniques, or pulling in of techniques from other schools of psychotherapy and re-evaluating these in terms of CBT models, is a necessary part of continued treatment development and provides one potentially very valuable way of increasing our treatments' effectiveness. However, this should be carried out with an awareness of the past work in order to avoid unnecessary repetition and waste, and re-branding or overcomplicating for its own sake should of course be avoided. Clinical innovation, whether from development of completely novel approaches or via re-evaluation of existing techniques, therefore represents an area where the interface of science and clinical practice is particularly important. New imagery techniques may be derived via a number of routes, for example via translation from basic research or via intuitive experimentation by clinicians in practice. However, in order to take such techniques forward, they should ideally be linked to mechanisms that can be articulated and tested in order to avoid a treatment being developed in an unhelpful direction simply because the initial putative mechanism was in fact not the one responsible for the beneficial effects (e.g. initial theories underlying the development of EMDR were almost certainly incorrect; Van den Hout and Engelhard 2012).
Given the ever-increasing number of apparently successful imagery-based techniques and CBT therapy approaches, one further clinical challenge for the future is the dissemination of these new approaches and their integration into routine clinical practice. If an imagery-focused therapy is developed for application in the context of a disorder for which there are already established CBT variants, and appears efficacious, what happens next? Should this new therapy be viewed as a replacement, an alternative option to choose from, or as providing a catalogue of ideas and techniques to be drawn on and integrated into the 'standard' approach? Perhaps we would wish to conduct a trial to compare this new treatment to existing ones, or to identify which treatment may be more indicated for which patients? However, such trials (particularly ones aiming to identify patient subgroups) require huge participant numbers, time, and therapist resources. Of course, this problem is not specific to new imagery-based approaches, but rather across the whole of CBT as researchers and clinicians continue to try to develop more effective therapies and more effective ways of tailoring therapy choices to individuals (see e.g. Dunn et al. 2019, for a broader perspective). One way forwards gaining increasing traction is to move away from 'brand-name' therapies and towards a focus on mechanisms and methods to target them (e.g. Hofmann and Hayes 2019; Holmes et al. 2018). The efficacy of an imagery-based approach, or incorporation of imagery into a technique, in targeting a specific mechanism could potentially be tested via experimental studies or 'micro-trials' amongst treatment-seeking individuals before incorporation into broader treatment frameworks, for example flexible 'modularised' therapy approaches (e.g. Black et al. 2018;Evans et al. 2020).

Implications for the Individual Researcher or Clinician
What does the current state of imagery in the science and practice of CBT imply for the ongoing work of individual researchers and clinicians? One clear implication is that clinicians and researchers need to have an awareness of mental imagery, its properties, and its potential dysfunctional manifestations in psychopathology, keeping it in mind when developing research concepts or conducting therapy. From a research perspective, given that imagery appears to be interwoven into so many cognitive and emotional processes, not considering whether or how imagery may play a role in any particular mechanism under investigation may result in an important facet of this mechanism being missed. From a clinical perspective, not asking about imagery risks missing out on important aspects of a patient's experience, and potentially key drivers of distress, impairment, or even risk (e.g. imagery of suicide or self-injury; Hales et al. 2011;Weßlau et al. 2015), and rules out the possibility of making use of simple and efficacious imagery-based treatment techniques. Or to frame this more positively, considering and asking about imagery opens up many opportunities for valuable clinical and research insights and a vast array of treatment possibilities. Further, understanding the basic science underlying imagery can lead to potential methods for augmenting 'standard' therapy techniques in a fairly uncomplicated manner (e.g. as illustrated in relation to thought records by Josefowitz 2017).
As a note of caution, there will of course be many individuals (both patients and therapists) who experience little imagery and do not find it a useful working tool in therapy. Further, given limited time and resources, principles of parsimony need to be applied-there are likely many situations in which using imagery-based techniques will not be necessary and may even distract from other, in specific cases more efficacious, aspects of therapies (especially if these are less attractive to therapists e.g. Waller 2009). Hence, delineating the boundaries of imagery's utility in therapy will also remain an important consideration. However, simply remembering to ask oneself "could there be a role for imagery here?" has great potential for opening up many avenues in both the science and practice of CBT.

Conclusions
Mental imagery has a long and varied history in the science and practice of CBT, and the recognition of its powerful effects has led to ever-increasing inclusion of imagery into theories informing treatment and into treatment protocols themselves. This leads to a current situation that is both full of exciting potential but also potentially overwhelming, given the now quite impressive breadth of the field. Challenges for future research and clinical practice include navigating the need to both go further and further into specific details, but at the same time to make trans-disciplinary connections and maintain an overview, retain the principles of parsimony in theory and clinical practice, and be on the lookout for limitations, counter-indications, and boundaries to the utility of imagery in treatment. Returning to basic principles of what exactly imagery is and how it occurs (to the best of our current understanding), and keeping a close eye on hypothesised mechanisms-and testing these wherever possible-can help to navigate this complexity. Both research and clinical outcomes indicate that imagery can have profound effects on emotions, cognitions, beliefs, and behaviour, and there is plenty of scope to improve our understanding of how best to harness or leverage this potential to effect therapeutic change in CBT.
Funding Open Access funding enabled and organized by Projekt DEAL.

Declarations
Conflict of Interest The author conducts and publishes research on mental imagery, and applies for grant funding for such research, but otherwise declares that he has no known conflicts of interest.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.