Background

A scientific theory can be defined as a well-substantiated explanation of some aspect of the natural world that is acquired through scientific methods and repeatedly tested and confirmed through observation and experimentation (Anonymous, 2014a ). Theories are animating and energising forces in research because they organise ideas and generate testable hypotheses. In medical research representations (models) of disease pathophysiology serve a similar purpose. In clinical practice, an understanding of pathophysiology is a component of the expertise that clinicians bring to their encounters with patients, helping to guide their decisions and facilitating communication about what has gone wrong when an illness occurs.

Although there are strong aetiological models for many diseases, psychiatric research has not yet established comprehensive scientific theories to explain the aetiology of depressive disorders, nor has a solid pathophysiological description emerged. This reality may reflect the complexity and multi-factorial aetiology of depressive disorders. Where detailed theories have emerged they have tended to focus on specific therapeutic modalities, such as the cognitive (Beck et al. Reference Beck, Rush, Shaw and Emery1979) or behavioural (Abramson et al. Reference Abramson, Seligman and Teasdale1978) formulations.

The term ‘biomedical model’ has been used to draw attention to an exclusion or devaluation of psychological and social factors in some aspects of psychiatric research (Anonymous, 2014b ). In distinction, social psychiatry can be defined as a ‘branch of psychiatry that focuses on the interpersonal and cultural context of mental disorder and mental wellbeing’ (Anonymous, 2014c ). Such models stand in contrast to the biopsychosocial model, a framework that encourages simultaneous attention to biological, psychological and social factors (Engel, Reference Engel1980). However, whereas all of these terms convey ideas about what may, or should be, considered important, they do not resemble more compelling types of theories or models that can organise and integrate factual observations and generate specific testable predictions. They do not provide, in themselves, a satisfying explanation of why mood disorders occur.

From outside of the field of medicine, the biomedical model is often perceived as a ‘dominant paradigm’ of psychiatric research (Bennett Johnson, Reference Bennett Johnson2012). However, the term is less frequently used by biomedical researchers. A Web of Science search for articles with the phrase ‘biomedical model’ in their title (search conducted on October 24, 2014) yielded only 60 papers. These had attracted (at that time), in aggregate, less than 600 citations. The most highly cited paper was concerned with swine genomics and the 2nd and 3rd most cited ones discussed breeding and laboratory techniques for opossums. In contrast, a similar search for titles containing the phrase ‘evidence-based medicine’ yielded nearly 3000 papers, attracting approximately 30 000 citations. Yet, the framework of evidence based medicine is not a theory or model. It is a general approach to clinical practice, one that favours the conscientious, explicit and judicious use of current best evidence in making decisions about the care of individual patients (Sackett et al. Reference Sackett, Rosenberg, Gray, Haynes and Richardson1996).

Some of the ideas that organise the thinking of depression researchers and animate their research activities may arise from informal models that are based, at least partially, on analogies to well-known disease processes. This idea cannot be decisively proven or disproven. However, it deserves to be articulated since such analogies may play a role in shaping the psychiatric knowledge-base. When models of depression are communicated to the public, they may tend to take the form of familiar disease-related scenarios. When decisions are made about research funding and acceptance of papers for publication, ideas that excite reviewers are likely to be the ones supported. Some such excitement may arise, at least partially, through models that arise as analogies or metaphors of familiar diseases. The goal of this paper is not to evaluate such models. The goal is simply to articulate these models.

Depression as a chemical imbalance

Many diseases are viewed as imbalances in allostatic or homeostatic physiological processes. Normal physiology includes countless examples of tightly regulated feedback loops and self-limiting processes. A good example is the regulation of thyroid and other endocrine hormones. A perturbation of regulatory mechanisms (e.g., due to a thyroid tumour that has escaped suppression or a thyroid gland unable to produce sufficient hormone despite maximal stimulation) leads to physiological imbalances that produce disease states (Larsen, Reference Larsen1982). Another example is Parkinson's disease, where motor symptoms are attributed to a dopamine–acetylcholine imbalance resulting from reduced striatal dopaminergic tone and subsequent cholinergic overactivity (Calabresi et al. Reference Calabresi, Picconi, Parnetti and Di Filippo2006). This general idea of an imbalance between physiological processes that are normally (in health) tightly regulated may contribute to the popularity of the common metaphor of a ‘chemical imbalance’. This particular metaphor has been criticised as representing something more akin to a cultural narrative than a theoretical driver of scientific thinking in psychiatry (France et al. Reference France, Lysaker and Robinson2007). It may be used by clinicians who feel that it reduces stigma (perhaps by diminishing any implication of moral responsibility for depression), although available evidence suggests that such explanations may actually be associated with increased stigma (Speerforck et al. Reference Speerforck, Schomerus, Pruess and Angermeyer2014).

Depression as a degenerative condition

Many familiar diseases are degenerative, involving decline or deterioration. A classic example is osteoarthritis which is often viewed as a ‘wear and tear’ condition that involves degenerative changes in joints such as the loss of articular cartilage, development of osteophytes, etc. (Bennell et al. Reference Bennell, Hunter and Hinman2012). As a concept, degeneration is an emotive term. It is quite alarming to think about a disease process that involves tissue destruction.

A few authors have directly posited an effect of ‘wear and tear’ on mental health, e.g. (Charles et al. Reference Charles, Piazza, Mogle, Sliwinski and Almeida2013). However, the concern most often expressed is that of neurotoxicity due to hypothalamic–pituitary–adrenal axis activation and leading to progressive processes associated with brain atrophy, particularly in the hippocampus (Sheline et al. Reference Sheline, Wang, Gado, Csernansky and Vannier1996). Such degenerative processes are also viewed as possible links between affective disorders and dementia (da Silva et al. Reference da Silva, Goncalves-Pereira, Xavier and Mukaetova-Ladinska2013). These ideas evoke powerful analogies. If anti-tumour necrosis factor biologic agents, for example, can prevent the progressive destruction of joints in rheumatoid arthritis (Taylor & Feldmann, Reference Taylor and Feldmann2009) can neuroprotective factors prevent the brain atrophy associated with depression (Boldrini et al. Reference Boldrini, Santiago, Hen, Dwork, Rosoklija, Tamir, Arango and John Mann2013; Ladea et al. Reference Ladea, Bran and Medrea2014)?

Depression as an outcome of toxicity

A particular perspective on the idea of degeneration in depressive disorders is the idea that abnormal development or degeneration may be precipitated by a toxic exposure. The poisoning metaphor is a compelling narrative, perhaps rendered even more so by analogy to the many types of poisoning that are encountered in medicine. Needleman, for example, described the ‘fourth stage’ understanding of lead toxicity in children as the recognition (starting in the 1970s) that lead-exposed children with no history of acute toxicity had deficits in IQ scores, attention, and language (Needleman, Reference Needleman2004). Similarly, a 2012 technical report from the American Academy of Pediatrics (Shonkoff et al. Reference Shonkoff and Garner2012) referred to ‘extensive evidence of the disruptive impacts of toxic stress, offering intriguing insights into causal mechanisms that link early adversity to later impairments in learning, behaviour, and both physical and mental well-being’. In alignment with the idea that stress can be a toxic exposure, depression was listed as an outcome of exposure to toxic stress, along with cognitive and language problems, similar long-accepted problems with lead exposure.

Depression as an injury

Although the Diagnostic and Statistical Manual of Mental Disorders (American Psychiatric Association, 2013) has a general philosophy of avoiding incorporation of judgments about aetiology into its diagnostic criteria, the incorporation of aetiology into some definitions is inevitable. Indeed, one of the new DSM-5 chapters is that for trauma- and stressor-related disorders (American Psychiatric Association, 2013). This type of cause–effect relationship resembles that of an injury – damage done by an external cause. The intuitive appeal of reducing long-term impacts by effective acute care for injuries, extended by analogy to post-transmatic stress discorder (PTSD), may have contributed to the rapid uptake and prolonged persistence of critical incident debriefing, despite inadequate evidence for its effectiveness (and evidence of possible harmfulness) (Rose & Bisson, Reference Rose and Bisson1998). The idea of an injury due to stress has been extended to the domain of depression, manifesting for example as a lasting effect of psychosocial trauma on the brain, with reduced left hippocampal volumes following traumatisation in women (Vythilingam et al. Reference Vythilingam, Heim, Newport, Miller, Anderson, Bronen, Brummer, Staib, Vermetten, Charney, Nemeroff and Bremner2002) or possible stress-sensitising (e.g., epigenetic or neurodevelopmental) effects of trauma (Taylor et al. Reference Taylor, Lerner, Sage, Lehman and Seeman2004; McLaughlin et al. Reference McLaughlin, Conron, Koenen and Gilman2010a , Reference McLaughlin, Kubzansky, Dunn, Waldinger, Vaillant and Koenen b ; Taylor, Reference Taylor2010).

Depression as a deficiency state

There are many familiar illnesses that result from a deficiency, either of a nutrient or of the internal production of a hormone. The idea that illnesses such as depression are due to a deficit in the production of a neurotransmitter has been, perhaps due to analogy, a very compelling idea. Within medical circles, this idea was popular in the 1960s and 1970s, with the monoamine theory of depression (Schildkraut, Reference Schildkraut1995). This model conceived of depression as a deficiency of one or more neurotransmitters. The metaphor was perhaps especially compelling as a result of the fact that many antidepressants increase synaptic levels of neurotransmitters involved in the regulation of mood. Another factor may have been the observation that mona-amine depleting drugs such as reserpine were purported to cause depression (Baumeister et al. Reference Baumeister, Hawkins and Uzelac2003). This deficiency model is now rarely put forward in the medical literature; however, an internet search using the search term ‘serotonin deficiency’ produces thousands of hits, suggesting that the idea has persisted. There continues to be occasional reference to this idea in the scientific literature as well (Angoa-Perez et al. Reference Angoa-Perez, Kane, Briggs, Herrera-Mundo, Sykes, Francescutti and Kuhn2014).

Depression as a progressive disease

The idea of a progressive disease is another compelling idea. It is one that suggests a need for urgent intervention. Cancer is perhaps the classical example. Malignancies begin at a microscopic level and are often difficult to detect early. They may be more difficult to treat after they have progressed to more advanced stages. The idea of a progressive disease lends credence to efforts at earlier detection, including the idea that detection should ideally occur in pre-symptomatic stages of a disease. A model of depression as a progressive disease may explain persisting interest in screening for depression, even though evidence for the effectiveness of this strategy has generally been lacking (Linden & Vodermaier, Reference Linden and Vodermaier2012; Thombs et al. Reference Thombs, Coyne, Cuijpers, de, Gilbody, Ioannidis, Johnson, Patten, Turner and Ziegelstein2012).

It has become generally accepted that the majority of common mental disorders have their onset during childhood or early adulthood, e.g., 75% by age 24 (Kessler et al. Reference Kessler, Berglund, Demler, Jin, Merikangas and Walters2005), which suggests that later presentations may represent advanced stages of a longstanding process. For example, Lin et al. (Reference Lin, Reniers and Wood2013) posited that an initially non-specific clinical picture for severe depression is followed by a worsening of symptoms and subsequently by acquisition of new symptoms that occur together with progressive neurobiological changes until a clearly recognisable disorder appears. The analogy to cancer is made even more explicit in a quotation from this paper: ‘the cancer analogy is useful here: minor surgery and local radiotherapy may be appropriate for early stages of breast cancer, whereas in later stages this would not be sufficient and more radical treatment such as mastectomy and chemotherapy might be indicated’ (Lin et al. Reference Lin, Reniers and Wood2013).

The idea of depression as a progressive disease may also augment a current perception of the need for biomarkers for depressive disorders. Biomarkers may precede clinical symptoms and thereby facilitate earlier intervention (Owens et al. Reference Owens, Herbert, Jones, Sahakian, Wilkinson, Dunn, Croudace and Goodyer2014).

Depression as an obsolete diagnostic term

There are examples in medicine of situations in which a primitive understanding of pathophysiological processes has led to inappropriate classification. Examples may be found in terms such as ‘dropsy’ or ‘consumption’ which became obsolete when more specific disease entities accounting for broadly defined clinical syndromes were described. Such conditions reflect diverse pathophysiological processes and their non-specific nature was an impediment to the selection of appropriate treatment and progress in research. If, by analogy, this applies to depressive disorders as currently defined, then deconstruction or elimination of existing categories may lead to more useful classifications. These historical scenarios resonate with discussions surrounding the emergence of alternative approaches to classification, such as RDoCs (Weinberger & Goldberg, Reference Weinberger and Goldberg2014).

Depressive disorders as mysterious conditions waiting for a breakthrough

The most well-known version of this metaphor is peptic ulcer disease and the discovery of helicobacter pylori. Prior to this discovery, the aetiology of this condition was poorly understood and a variety of poorly replicated associations with aetiological factors had been reported. Everything changed with the discovery of this, previously unknown, determinant (Marshall & Warren, Reference Marshall and Warren1984). Similarly, the discovery of the role of human pappiloma virus in cervical cancer was a ‘game changing’ breakthrough (Walboomers et al. Reference Walboomers, Jacobs, Manos, Bosch, Kummer, Shah, Snijders, Peto, Meijer and Munoz1999). This virus is now regarded as a necessary case of most cervical cancers, guiding the field in a whole new direction for prevention (primary prevention through vaccination as opposed to reliance on secondary prevention through Pap tests). The idea that there may be a hidden secret to the aetiology of poorly understood conditions such as depressive disorders is perhaps made more compelling by these examples. Perhaps the best evidence of the power of this idea is the great interest in the microbiome as a determinant of depression and other common disorders (Foster & Neufeld, Reference Foster and Neufeld2013). There is naturally a great excitement attached to the idea that a ‘game changing’ discovery might completely reorganise thinking about a mysterious condition such as depression.

Depression as an evolutionary vestige

Evolutionary models of depression are usually based on the idea that symptoms such as sadness or loss of interest have an adaptive role and that depressive disorders arise as dysregulated or malignant version of this, evolutionarily adaptive, response (Hagen, Reference Hagen2011). The adaptive mechanisms may include displays of emotion that reinforce social or parental connections, act as involuntary subordination strategies arising during social competition, or that may play a role analogous to physical pain in motivating avoidance of harmful situations (Hagen, Reference Hagen2011). Such strategies may have had an adaptive role during evolutionary history but may more often be maladaptive in the modern context. As with the other examples listed in this commentary, this concept aligns with other examples of disease pathophysiology. For example, the caecal appendix has often been regarded an evolutionary vestige that is nevertheless capable of producing disease. Recent insights into the function of the appendix as a well-adapted maintainer of mutualistic intestinal flora during intestinal infections (a ‘safe house’) have not challenged the idea that this function is largely obsolete in post-industrialised settings (Laurin et al. Reference Laurin, Everett and Parker2011). Related concepts may be found in the asthma and allergy literature, where the hygiene hypothesis posits that a mismatch between evolutionarily determined aspects of immune responsiveness and modern living environments may be partially responsible for epidemics of such illnesses (Eder et al. Reference Eder, Ege and von Multius2006; Brooks et al. Reference Brooks, Pearce and Douwes2013). Indeed, the hygiene hypothesis and its associated implications for inflammatory regulation have been invoked with reference to the aetiology of depressive disorders (Raison & Miller, Reference Raison and Miller2013). Another example is obesity since adiposity may be understood as reflecting a ‘thrifty genotype’ that was adaptive in the evolutionary past or as a complex risk management system poorly adapted to contemporary social and economic environments (Wells, Reference Wells2012).