Imaging cancer: image-based diagnostic communication in radiologists’ embodied cognition

Introduction

This article presents some key findings based on ethnographic observations of oncoradiologists at work (hidden for anonymity). ¹ Careful observation of this specific diagnostic praxis shows medical images as tools, instructions for action or navigation maps, rather than non-contextual static depictions. In doing so, this approach sheds new light on how the technological enchantment of images helps to communicate diagnosis and activate relevant professional skills. To get to the answers, it is first necessary to establish what I understand as the ‘enactive approach’ ² to entail and investigate how it may enter the image-based medical diagnostic praxis. The key influence is Shaun Gallagher’s suggestion of a ‘front-loaded’ phenomenology: ‘The idea is to front load phenomenological insights into the design of experiments, that is, to allow the insights developed in phenomenological analyses . . . to inform the way experiments are set up’ (Gallagher, 2003: 91). I believe it is equally possible to front load phenomenological ideas via situated ethnographic research as it is through the experimentally conditioned environment. Indeed, Gallagher himself admits that there are problems related to the artificiality of experimental conditions and the enactive approach. Situating phenomenological and enactive concepts (for example, categorial intuition, sense of authorship of one’s actions or the experience of feedback while performing a task) suggests some new perspectives for observing professional action and presupposes practical implications. For example, in the case of oncoradiology, triadic structure of image apprehension can suggest novel visual search patterns and tips for the improvement of imaging software and hardware.

While the enactive approach to cognition borrowed much from phenomenology, it eventually became an autonomous, albeit non-monolithic, movement. ³ However, all enactivist theories in one way or another ground cognition in the capacities for dynamic interaction between organism and environment, which is only possible for an embodied being. This not only presupposes an original deconstruction of traditional philosophical dichotomies, such as the subject–object dualism, or consciousness as a local brain property, but also opens the way for socially relevant applications. Experience here is not a neutral contemplation but an effort to use and/or change the given and consciousness embodied and embedded in the physical and social world (e.g. Hutto and Myin, 2013, 2017; Noë, 2006; O’Regan and Noë, 2001; Varela et al., 1991). A radiologist’s ‘body schema’ (Gallagher, 1986, 2005; Merleau-Ponty, 1962) is not a static representation, but a ‘dynamic’ capacity, inasmuch as this body appears to the radiologist as a posture with a view to a certain actual or possible task. Importantly, such ‘sensorimotor understanding’ is construed as implicit and practical, rather than explicit and propositional (Gallagher, 2005; Noë, 2004). Because professionally relevant experiences are embodied, embedded and enacted in the world, it is difficult to research the lived aspects of cognition in scientific labs. That said, an enactive approach enables us to thematize the whole process and to document those units which are usually of little or no interest to theoretical or experimental approaches; for example, by giving very detailed descriptions of radiologists’ bodily movements, their skilful anticipatory modalities, coping with technological assistance, communication by language and gestures, etc.

Enactivism comprises a set of theories that approach cognition in an alternative way to both contemporary science and traditional cognitivism. Thus, the genesis of knowledge is seen to be based on interaction, that is, various practices – for example, medicine (Mol, 2002), which helps to situate enactive investigations in the living context of social praxis. Furthermore, enactivism stresses the role of the encounter between an organism and its environment for the origins of cognition as well as its embodied, action-oriented, pre-reflective, and even social nature. In that sense, enactivism is strongly anti-representationalist (Gallagher, 2017).

By adopting an enactivist approach, this article demonstrates that the radiologist’s experiential accomplishments, as well as the whole diagnostic process, have their origins in a rather a-subjective domain of environmental, technologically mediated, intersubjective, institutional and sensorial relations. This is why this particular piece of research has no intention of experimenting (Latour, 1987) with radiologists (as in, for instance, Drew et al., 2013).

It is true that there are already many studies of scientific imaging grounded in empirical analyses of practising scientists but, despite some affinities, those other analyses draw from different theoretical perspectives to analyse imaging practices than the enactive approach presented here. First of all, this research may be seen as an example of the social and cognitive organization of a profession (Goodwin, 1994). However, it narrows down the medical profession to image-based onco-diagnostics and takes the form of situated, anti-representational cognition (see Mol, 2002).

Major situated cognition alternatives to the analysis of medical imaging based on the enactive approach are semiotics and postphenomenology. Fascinating conceptual and practical research has been carried out in these fields in relation to medical imaging. However, the enactivist perspective brings to light something that others could not or cannot do. For example, in the case of semiotic analysis of how fMRI researchers actively involve their bodies (Alač, 2011), the relation between constitution and communication of signs and body-schematic coupling with the image-mediated environment is somehow reductive (for example, concentrating on hand movement), while enactivism is concerned with the involvement of the ‘whole’ body (body schema) in the experience, for example, of vision. Furthermore, as other semiotics join the discussion (Baxter et al., 2018; Kergosien, 1991), no attention is given to such enactive projects as the inner and external horizons of imaging or the affective pressure that imaging technologies place on radiologists; instead, the focus is more on the radiologist’s skilful actions and conversations, but not its embodied origins. Postphenomenology, conversely, tends to lean on the importance of a technological medium (Fried and Rosenberger, 2021) thus risking the disruption of the fragile balance of enactive strategies performed by individuals and the affordances provided by media ecologies which is so characteristic of enactivism. Technological mediation, according to postphenomenologists, means that scientific instruments actively mediate how reality becomes present to scientists (De Boer et al., 2021), undermining the autonomy of meaningful performances by autopoietic systems (Maturana and Varela, 1980). Even in clinical discussions of the topic, the focus is usually placed on decision making rather than its embodied origins, although the non-individual character of the latter is acknowledged (Van Baalen et al., 2017). Finally, most of these perspectives do not make explicit the specificities of cancer, which makes oncoradiology stand out even in relation to other spheres of medical imaging.

Because of its stress on active embodied engagement with the practical environment by humans as biological and social beings (De Jaegher et al., 2010), the role of the body in enactive phenomenology is extended to the constitution of higher objectivities – the constitution of values. Second, if we follow the major dictum of enactivism, i.e. that there is no gap between action and cognition, then we witness an enormous expansion of the possibilities to routinely set in motion (enact) certain necessary experiences, for example, via imaging technology (Noë, 2012). This bringing together of the enactive approach to cognition and habitual coping with technology constitutes the possibility of renewing certain classic concepts of phenomenology and enactivism (image-consciousness, feedback, affordances, modification, neutralization, etc.). In this way, meanings so fundamental to classic phenomenology became rules for use and interpretation at the same time, which also presupposes a possible practical improvement of them. And the thematization, as such, made concrete the horizons (see later) of usage, including in professions (such as radiology, where the radiologist emerges as a practical problem-solver performing context-embedded actions via image-based technology which mediates and amplifies poles of practical primacies). The professional environment thus serves as an ‘outside’ memory, i.e. an external information store which can be accessed as needed through skilful movements (Noë, 2006; Ryle, 1962; Stiegler, 1998). Therefore, technology mediates not only perception but also motor skills and feedback affection (see later).

Importantly, such ‘sensorimotor understanding’ is construed as implicit and practical, rather than explicit and propositional (Dreyfus, 2014; Gallagher, 2005; Noë, 2006). Because professionally relevant experiences are embodied, embedded and enacted in the world, it makes it difficult to research the lived aspects of cognition in scientific labs. That said, an enactive approach enables us to thematize the whole process and to document those units which are usually of little or no interest to theoretical or experimental approaches. Hence, the article does not aim to develop a comprehensive theory of radiology imaging, but rather focuses on specific moments of oncoradiology praxis. These moments include both affection by radiology imaging (sections on non-representational and amodal character of oncoradiology imaging, reduction to essential parts, etc.) and some enactive strategies usually performed by the radiologist (switching apprehensions and modes of imaging, the importance of professional feedback, grasp of a categorial situation, usage of measuring software, etc.). Instead of detailed ethnographic vignettes (Briedis, 2019, 2020), in order to support the arguments given in those sections, some actual radiograms are presented and interpreted from the enactive point of view.

1 Unique Features Of Oncoradiology Imaging

Despite radiology’s diagnostic speed (Tabár et al., 2005), the radiologist’s motivational structure is not that of a mindless reflex or inference, but rather presupposes a complex cognition. Pure passive perception is simply not possible for the human mind and seeing is never merely registering. As such, experience is not a neutral contemplation but an effort to use and/or change the given and consciousness embodied and embedded in the physical and social world (e.g. Hutto and Myin, 2013, 2017; Noë, 2006; O’Regan and Noë, 2001; Varela et al., 1991). Therefore, the radiologist’s visual experiences can be seen as skilful acting, as (intersubjectively) mediated knowledge of bodily (sensorimotor) transformations. Moreover, here, the seemingly ‘personal’ professional experiences of the radiologist are constituted out of the history of encounters with a materially and socially established priority of things, especially things mediated by specific technology. ⁴

In oncoradiology, what is imaged is often not enough for the familiar form to come forward. There are some analogical moments (where there is some resemblance between the subject and object of an image), but mostly non-analogical moments, wherein image presentation points beyond the image. If a silhouette deviates significantly, then the image not so much represents as gives meaning to the subject. This meaning presupposes that a skilful subject, i.e. the radiologist who enacts, brings forth meaning in action.

Some technology can better assist in bringing out this meaning, even concerning non-intentional moments of experience (feedback, affordances, neutralization, etc., see e.g., De Boer et al., 2021). Hence, if we view the foundations of the radiologist’s diagnostic performance as enaction, that is, setting in motion a certain experience via technology and routine embodiment, we can analyse the pre-reflective although meaningful relation between radiologists and their environment shaped by specific imaging technologies. This hypothesis is based on the assumption that pictures are devices that can constitute cognitive processes and not mere representations. Secondly, devices create the conditions of possibility for a specific cognitive act: if human beings enact pictures, then the pictures themselves must be conceived of as physical constitutive extensions (see further) of some aspects of our cognition.

Cancer imaging is special because cancer detection rests on more abstract and minimalist images, which redirect and complete experienced radiologists’ attention from the image given to the ‘missing parts’ of the pathological process. So, for example, the radiologist can grasp the oncological pathology as the connection between a primary tumour and metastasis which may be situated in different (and not visually presented) areas, but its location would still depend on the blood flow and other biological regularities. This means that the radiologist more imagines than perceives ⁵ (a novice sees imaged features, while an experienced radiologist by way of imagination brings diagnosis to the exact pathology, differentiates it, anticipates physical and social impairment, for example, limited movements or changes in eating, dressing, working or sexual routine and makes a prognosis) based on the knowledge of qualitative causality which is characteristic of cancer. The situation with cancer imaging also demands more skilful action (choosing the imaging modality, rearranging images, navigating, evaluating image quality, changing perspectives, communicating with clinical data, etc.) than just a static experience of depictive representation (see below). The non-mimetic character of oncoradiology imaging suggests that it is an active part of a subject’s (cancer) constitution rather than a passive mirroring of objective reality because it depends on the radiologist’s diagnostic task. A surgeon, for example, would have a completely different perspective, image and task in respect of that same tumour (not a depiction of it).

In addition to this anti-representational aspect, there are other features that are unique to oncoradiology, which taken together will attest that this technologically advanced imaging helps to bring forth certain phenomena (causality, ⁶ absent parts, movement, temporality, among others) better than other intentions, such as perception, language or free imagination (Husserl, 2005[1950 onwards]).

2 The Embodied Origins Of The Radiologist’s Skilful Action And Professional Feedback

Despite these linguistic examples, we must not forget that while radiology imaging captures causality in many ways, it is the radiologist’s body that brings relevant causal relations to apprehension. The rejection of images as representational in oncoradiology goes with the acceptance of the body as the original foundation of the radiologist’s mind. Moreover, this body is also not a static representation, but a body schema; that is, it has a ‘dynamic’ capacity, inasmuch as it transparently renders the possibility to engage in certain actual or possible tasks without appearing to the subject. This body explains how decisions are not made by causes or inferences; instead, the body oscillates through various degrees of mindfulness and is motivated by both internal (what is on the screen) and external (the social values of a given praxis) horizons of medical imaging. Acts of consciousness cannot have the cause, but they are motivated. Motive and action are linked internally, while cause and effect are linked empirically. Empirical obstacles can prevent action for diagnosis, but the desire to act is already there in the radiologist’s motivation. It is crucial that this desire is not an internal event and there is no internal diagnostically represented goal without presupposing action, which in turn is founded on the affordances to execute such action and a certain self-affection.

Every professional has a kind of immediate knowledge about how to understand their environment, i.e. their ‘typical’ (Schutz, 1967) constants of embodied activity. It is important that in radiology this typification of an environment has its social as well as perceptual aspects. Work is the object and framework of my actions at once, while being pragmatic and organized into the hierarchy of graspable zones, the closest of which Schutz calls the ‘manipulatory zone’ with its own spatiotemporal horizon (modified by imaging). This zone in turn is constituted by my constant ‘I can do it again’ idealization (radiology is very repetitive). After evident, fulfilled judgment, habitual conviction is constituted with its categorial bodily aspects. The person lives in the conviction of that judgment and constantly returns to the understanding of the sense of that object. Every successful fulfillment leaves a trace in the ego and turns into a habit, which has a form of ‘empty’ practical possessing and social habituality (norms), not static representational memory.

The person develops a ‘habitual’ style of thinking, feeling, willing and acting – what person and professional he or she is. Ego is stabilized via habits (as in ‘I can’). Habit in this sense organizes experiences into the horizons of familiarity and unfamiliarity, where relevant appearances have their distinct style of appearing (Noë, 2012). There are intellectual habits as well. So, you may have not only daily but also professional habits, especially if your profession is connected to the scientific worldview, as is the case with medicine. Then, for example, the natural attitude might have very powerful habitual drives, which predispose your proprioception, temporality, attitude toward physical objects, etc.

Moreover, the multimodal experience of categorial situations, especially their absent parts (indeterminants), is inseparable from a certain pre-reflective monitoring of how things are going for you and what affordances (Gibson, 1950) are out there for your next move. Hence, part of cognition is to anticipate the affective tone of a certain object and/or action. It also requires an attention to feedback about what is concealed. Open intention means that the radiologist knows about concealed parts and decides on enaction or inaction, because cognition, as it were, is also about knowing what not to do. It is peculiar that this experiential feedback somehow gives information but does not represent (an object, task, judgment or proposition) and, as such, it is the genesis of autonomous professional action.

Such phenomenological structures as the urge for maximal grip (Merleau-Ponty, 1962) show that perception is, in its own way, normative. Perception forms a significant part of the visual experience, which involves navigating through the region while oscillating between descriptive features and normative ones. This also concerns consciousness and self-reflection – we not only perceive something but also ‘know’ how well we perceive it. This leads to new actions. After non-conceptual perception discloses the world as a familiar space, this non-conceptual givenness is converted into conceptual content so that perception can fulfill certain empty intentions according to the regional task (type of pathology and mode of diagnosis).

In spite of the apparent fact that the radiologist deals with successions of images, they need to be treated as particular regimes of variation which is not the same as a stable succession – for example, when moving from a neutral image to the disclosure of pathology. It is crucial that, besides imaged contents, the radiologist’s feeling of the power to act during these successions is also decisive. To be a radiologist, from the enactive point of view, then, means to experience a continuation of a particular variation, which in turn depends on multiple factors, such as overall experience, personalized use of imaging technology, knowledge of the particular mode and communication with colleagues. And the affect is not reducible to an intellectual comparison of ideas, but is constituted by the lived transition or lived passage from one degree to another. The radiologist follows such affects which increase the power of acting. Novice radiologists follow the gestural guidance of trainers, gradually supplementing this affection by guidance with the autonomous experience of authorship of their diagnostic actions.

Of course, affection (knowledge without representing) is just the first step in grasping the causal category and moving accordingly, while egoic, propositional levels follow. The radiologist then equals the ability for certain movements, action. In the diagnostic task, the radiologist forms an individuality on the level of embodiment such that the relations between the radiologist’s body parts (especially hand–eye–voice coordination as presupposing some common strata of motivated embodiment), the rhythm of movement (arrangement of image sequences and image scanning) and rest (unattended images and imaged areas) are always in danger of being destroyed by the dead end of affordances which happens frequently to novice radiologists. This might result in a mistake, the need to change strategy, repeating the positioning, or consulting a colleague, but it originates in a certain destruction of the whole bodily composition.

Conclusions

What makes some features and/or processes given via radiological image ‘look’ abnormal is not a depiction of the ‘objective’ state of affairs, but rather enactment: the engagement in a particular culturally shaped perceptual process. From the enactivist perspective, the professional skills of a radiologist may be measured by his or her ability to manipulate various ways of the disclosure of diagnostic meaning. This includes not only the enactive manipulation of imaging technology, but also openness to the specific affection by radiology images. Radiologists’ diagnostic environment is affected by various modes of medical imaging and regional knowledge (the social normativity of basic abnormal progressions). Cognitively, this founds strategies of embodied cognition and presupposes the ability to manipulate several apprehensions based on the same sensations. ‘Presence’ does not assume priority here, as a number of absent, non-imaged constituents of a pathological situation influence the genesis of diagnosis. Furthermore, radiology images are not representations in a traditional depictive way and eventually there is no internal represented diagnostic goal ‘in the radiologist’s head’ without presupposing skilful action, which in turn includes pre-reflective cognition of professional affordances and a certain degree of self-consciousness as a feedback experience on performing professional tasks.

The traces of this embodied process can also be found in the language that radiologists use while interpreting and communicating causal relations which are modified by radiology imaging, for example, appearing as stripped back to their essential features and prepared for various strategies of image navigation. Hence, modification by radiology imaging constitutes a form of prosthesis for the radiologist’s embodied cognition, that is, a field of possible strategies of image navigation for which the design of imaging technology software and hardware is crucial. Hence, the enactivist perspective may have some practical implications for oncoradiology, for example, in the organization of working surroundings according to radiologists’ embodied strategies, the improvement of diagnostic software, and the classification of pathologies, all of which can give rise to new practices of teaching (intersubjective relation between novice and experienced radiologists) that are different from those grounded in a representationalist perspective. In short, an enactive take on oncoradiology imaging suggests a number of alternatives to the traditional memorizing–retrieval model of diagnostics, which constitutes a very promising field for future research.