Mechanistic Silence, Clinical Efficacy: Rethinking the Epistemic Standards We Apply to Mental Health Treatment

Abstract

A common objection to psychiatric and psychological treatment holds that because the brain mechanisms underlying disorders such as depression and anxiety are incompletely understood, the treatments aimed at them cannot be trusted. This essay argues that the objection rests on a conflation of two distinct epistemic projects—establishing that a treatment works and explaining why it works—that are supported by different kinds of evidence and that historically progress on independent timelines. Drawing on the histories of aspirin, general anesthesia, and lithium, three interventions whose mechanisms remained unresolved for decades or are still debated, the essay shows that mechanistic uncertainty is the medical norm rather than a psychiatric exception. It then examines the specific case of antidepressants and the contested serotonin hypothesis, arguing that recent controversy over that hypothesis speaks to the fragility of one biochemical narrative, not to the empirical evidence for drug efficacy itself. Finally, the essay disaggregates "mental health treatment" into pharmacological and psychotherapeutic components, showing that evidence-based psychotherapies operate through comparatively well-characterized psychological processes and perform comparably to, or better than, medication for mild-to-moderate presentations. The essay concludes that skepticism about mental health care should be aimed precisely at the claims the evidence does not support, rather than at the field as a whole.

1. Introduction

Skepticism toward psychiatry frequently takes a particular form: because clinicians and researchers cannot fully specify the neurobiological chain of events by which a given treatment relieves depression or anxiety, the treatment itself is treated as suspect, or the field is characterized as having failed to establish a scientific basis for its interventions. This argument has intuitive appeal. It seems reasonable to expect that a legitimate medical treatment should come with an explanation of its mechanism, and the absence of such an explanation can feel like an admission that clinicians do not actually understand what they are doing.

The argument, however, depends on a premise that does not hold up under scrutiny: that mechanistic understanding and demonstrated efficacy are the same kind of claim, established by the same kind of evidence, and that a gap between them indicates failure rather than the ordinary condition of applied medicine. This essay examines that premise directly. It draws on cases from general medicine in which treatments were adopted, refined, and relied upon for decades before their mechanisms were elucidated—and in some cases remain a matter of live scientific dispute—to argue that the mechanistic gap in psychiatry is a difference of degree, not of kind, from the rest of clinical medicine. It then turns to the specific controversy over antidepressants and serotonin, and finally to the distinction between pharmacological and psychotherapeutic treatment, to give a more precise account of where legitimate skepticism is warranted and where it is not.

2. Two Kinds of Evidence: Efficacy and Mechanism

Modern evidence-based medicine (EBM) is built on a hierarchy that places randomized controlled trials (RCTs) and their meta-analytic syntheses near the top of the evidentiary pyramid, while mechanistic or pathophysiological reasoning is traditionally placed near the bottom (Howick, 2011). This ordering is often misread as a claim that mechanisms do not matter. The more accurate reading is methodological: RCTs are designed to answer a narrow but well-controlled causal question—does this intervention, on average, produce a better outcome than the comparator, in this population?—by using randomization to neutralize confounding variables that would otherwise make cause and effect difficult to disentangle. Mechanistic research answers a different question entirely: through which biological, cellular, or molecular pathway does the intervention produce its effect? Howick (2011) has argued that only inferences drawn from mechanisms, rather than mechanisms as such, count as evidence, and that mechanistic reasoning is not required to establish causation once high-quality controlled studies are available, even though mechanistic reasoning is also more problematic as a standalone evidentiary tool than is often assumed.

The practical consequence is that a field can be simultaneously mature on the efficacy question and immature on the mechanism question, because the two tracks of research require different tools, different timescales, and often different scientific communities entirely. Clinical trial infrastructure, statistical methodology, and regulatory approval pathways have been built specifically to answer the "does it work" question at scale. Basic neuroscience, molecular biology, and pharmacology are comparatively slower and more exploratory disciplines, particularly when the target organ—the human brain—is largely inaccessible to direct study in living patients. This is not a psychiatric peculiarity; it is a structural feature of how biomedical knowledge accumulates.

3. Mechanistic Silence Is the Medical Norm

3.1 Aspirin

Aspirin (acetylsalicylic acid) was synthesized and marketed by Bayer in 1897 and became, in short order, one of the most widely used medicines in history, prescribed for pain, fever, and inflammation on the strength of clinical observation alone. Its mechanism of action remained unknown for seventy-four years, until John Vane demonstrated in 1971 that aspirin and related compounds inhibited the synthesis of prostaglandins, in proportion to their anti-inflammatory potency in humans. Vane's discovery, later recognized with a Nobel Prize, showed that aspirin blocks the cyclooxygenase enzyme responsible for prostaglandin production, which in turn explained both its therapeutic effects and its characteristic side effects, such as gastric irritation (Vane & Botting, 2003). For nearly three-quarters of a century, then, physicians correctly and confidently prescribed one of the world's most common drugs without any empirically validated account of how it worked. The absence of mechanistic knowledge did not prevent aspirin from being effective, nor did it prevent clinicians from using it appropriately based on accumulated clinical experience and, later, controlled trials.

3.2 General Anesthesia

General anesthesia presents an even more striking case, because the mechanistic gap has not closed. Anesthetics produce widespread neurodepression in the central nervous system by enhancing inhibitory neurotransmission and reducing excitatory neurotransmission, yet the precise action mechanisms are not completely understood, and the anesthetic state itself comprises multiple components—amnesia, unconsciousness, analgesia, and immobility—each apparently mediated by different receptors and neuronal pathways (Son, 2010). More than 170 years after the first surgical use of ether, and with more than 300 million patients undergoing general anesthesia annually, the field still lacks a unified account of how these drugs produce loss of consciousness, and recent reviews continue to describe the underlying neural network mechanisms as poorly understood despite considerable technological advances in neuroscience methodology. General anesthesia is arguably the single most consequential and routinely used intervention in modern surgical medicine, and it is administered daily worldwide by clinicians who cannot fully explain, at a mechanistic level, how it achieves its primary effect.

3.3 Lithium

Lithium occupies a similar position within psychiatry specifically, which makes it a useful bridge case. Introduced into clinical psychiatry by John Cade in 1949, lithium remains one of the most effective treatments ever discovered for bipolar disorder, including a well-documented reduction in suicide risk. Despite lithium's presence in modern psychiatry for three-quarters of a century, its mechanisms of therapeutic action have not been fully elucidated, and researchers have proposed and revisited a wide range of candidate pathways, including effects on second messenger systems, glycogen synthase kinase-3β, neurotrophic factors, immune function, biological rhythms, and mitochondrial processes, without converging on a single definitive account (Sakrajda & Rybakowski, 2025). Contemporary neuroimaging and gene-expression studies have added further candidate mechanisms without resolving the underlying uncertainty (Anand et al., 2020). Lithium is thus a live demonstration, from within psychiatry itself, that a treatment can be both highly effective and mechanistically opaque at the same time.

Taken together, these three cases establish a general pattern: mechanistic elucidation in medicine typically lags—sometimes by decades, sometimes indefinitely—behind the establishment of clinical efficacy. This pattern holds across surgical anesthesiology, general pharmacology, and psychiatry alike. If the absence of a settled mechanism were sufficient grounds for distrust, virtually no domain of clinical medicine would survive the standard, aspirin and general anesthesia included.

4. The Specific Case of Antidepressants and Serotonin

The mechanistic uncertainty surrounding antidepressants deserves separate treatment, because it has become entangled in recent years with a public controversy that is frequently, and inaccurately, read as evidence that antidepressants do not work. In 2022, Moncrieff and colleagues published a systematic umbrella review in Molecular Psychiatry concluding that the principal areas of serotonin research—including serotonin and its metabolites in body fluids, receptor binding, transporter levels, tryptophan depletion studies, and gene associations—provided no consistent evidence that depression is associated with lowered serotonin concentration or activity. The review received enormous public attention and was widely interpreted, including by the authors themselves in public commentary, as undermining the rationale for prescribing antidepressants.

That interpretation did not go unchallenged within the scientific community. A response authored by thirty-six researchers and led by Sameer Jauhar of King's College London, titled "A leaky umbrella has little value," argued that the original review was methodologically inconsistent with standard umbrella-review practice, since its authors summarized existing results rather than independently extracting and reanalyzing the underlying data, and that it selectively drew on individual studies in ways that a properly conducted umbrella review is designed to avoid (Jauhar et al., 2023). The exchange illustrates a genuine and unresolved scientific dispute about a specific biochemical hypothesis—but it is important to see precisely what is, and is not, at stake in that dispute. The serotonin hypothesis was always intended as one candidate explanation of how certain antidepressants might work, not as the evidentiary basis for whether they work. Those two questions are answered by different literatures. Independent patient-level meta-analyses have found that antidepressant medication produces a modest but statistically reliable advantage over cognitive behavioral therapy in reducing depression severity (Boschloo et al., 2019), and large network meta-analyses comparing antidepressants against placebo across dozens of trials have established that these drugs outperform placebo for major depressive disorder, even though the biochemical account of why they do so remains contested (Cipriani et al., 2018). A collapse in confidence about one candidate mechanism is not equivalent to, and does not require, a collapse in confidence about the treatment's demonstrated clinical effect. Conflating the two is precisely the error this essay is concerned with.

5. Disaggregating "Mental Health Treatment"

A further source of confusion in popular skepticism is the tendency to treat "mental health treatment" and "psychiatric medication" as synonyms. They are not, and the distinction matters for exactly the argument developed above. Evidence-based psychotherapies, and particularly cognitive behavioral therapy (CBT), operate through processes that are considerably more accessible to direct empirical study than the molecular pharmacology of psychotropic drugs, because their proposed mechanisms—changes in maladaptive cognitions, exposure-based extinction of avoidance behavior, behavioral activation—are psychological constructs that can be measured within the course of treatment itself, rather than requiring invasive access to brain tissue.

This does not mean the mechanistic picture for psychotherapy is fully resolved; it is more candid to say that the uncertainty there takes a different and, in some ways, more tractable form. Kazdin (2007) observed that after decades of psychotherapy research culminating in the identification of several treatments with strong efficacy evidence, the field still could not offer an evidence-based explanation of how or why even its most thoroughly studied interventions produce change—that is, the field faced the same efficacy–mechanism gap seen elsewhere in medicine, arising in part because establishing that a therapy is effective does not, by itself, establish which of its components are the active mediating ingredients. What distinguishes psychotherapy research is that the tools for testing candidate mechanisms—session-by-session tracking of cognitive change, behavioral avoidance, and therapeutic alliance—are more directly observable and manipulable in a clinical or laboratory setting than are the intracellular signaling cascades implicated in pharmacological treatment, even if a fully validated causal account remains elusive in both domains.

On the efficacy side, the evidence for psychotherapy in mild-to-moderate depression and anxiety is substantial. Meta-analytic estimates place CBT's effect size for major depression in a range comparable to antidepressant medication, with CBT showing a moderate-to-large advantage over placebo conditions and demonstrating durability of benefit that reduces relapse risk after treatment ends, while medication tends to produce a somewhat faster initial symptom response (see, e.g., the comparative findings summarized above; Boschloo et al., 2019). This body of evidence supports a more precise object for skepticism than "mental health treatment" as an undifferentiated category. A model of care built around a rushed medication-only consultation is a reasonable target for critique on grounds of quality and accessibility. A model of care that offers structured psychotherapy, informed medication management where indicated, and attention to lifestyle and psychosocial factors is a different clinical entity, addressing a different empirical base, and it does not inherit the weaknesses of the narrower model merely by sharing a name with it.

6. Implications

Three implications follow from the argument developed here. First, the demand that a treatment demonstrate its mechanism before being trusted is not a neutral scientific standard; it is a standard that, if applied consistently, would disqualify substantial portions of accepted medical practice, including some of its most successful interventions. Consistency requires either applying the same skepticism to aspirin and general anesthesia, which is not typically done, or recognizing that mechanistic uncertainty alone does not settle the question of clinical trust.

Second, public debates over specific biochemical hypotheses, such as the serotonin theory of depression, should be evaluated on their own terms as contributions to mechanistic understanding, and should not be treated as referenda on treatment efficacy, which rests on a separate and more robust evidentiary base drawn from randomized trials and meta-analyses.

Third, more precise public and clinical communication would distinguish between psychiatric medication, psychotherapy, and the broader configuration of care in which either is delivered. Skepticism aimed at a specific delivery model—for instance, brief medication-only consultations without adequate follow-up—identifies a genuine and addressable problem in how care is often organized. Skepticism aimed at the underlying evidence base for treatment efficacy is a different and substantially weaker claim, not well supported by the literature reviewed here.

7. Conclusion

The absence of a fully elucidated mechanism is not a psychiatric failing; it is the ordinary condition of medicine, visible in the history of aspirin, in the continuing uncertainty surrounding general anesthesia, and within psychiatry's own use of lithium. The recent controversy over the serotonin hypothesis of depression is best understood as an unsettled dispute about one candidate biological explanation, not as evidence against the substantial trial-based literature supporting antidepressant efficacy. And the common practice of treating "mental health treatment" as synonymous with medication overlooks a large and comparatively well-supported psychotherapy literature with its own, more tractable, evidence on mechanisms of change. A more defensible form of skepticism would target specific, falsifiable claims—about a particular biochemical hypothesis, or about a particular model of care delivery—rather than the field's evidentiary foundations as a whole.

References

Anand, A., Nakamura, K., Spielberg, J. M., Cha, J., Karne, H., & Hu, B. (2020). Integrative analysis of lithium treatment associated effects on brain structure and peripheral gene expression reveals novel molecular insights into mechanism of action. Translational Psychiatry, 10(1), 103. https://doi.org/10.1038/s41398-020-0784-z

Boschloo, L., Bekhuis, E., Weitz, E. S., Reijnders, M., DeRubeis, R. J., Dimidjian, S., Dunlop, B. W., Fournier, J. C., Hegerl, U., Hollon, S. D., Jarrett, R. B., Kennedy, S. H., Miranda, J., Mohr, D. C., Simons, A. D., Vittengl, J. R., Cuijpers, P., & Schoevers, R. A. (2019). The symptom-specific efficacy of antidepressant medication vs. cognitive behavioral therapy in the treatment of depression: Results from an individual patient data meta-analysis. World Psychiatry, 18(2), 183–191. https://doi.org/10.1002/wps.20630

Cipriani, A., Furukawa, T. A., Salanti, G., Chaimani, A., Atkinson, L. Z., Ogawa, Y., Leucht, S., Ruhe, H. G., Turner, E. H., Higgins, J. P. T., Egger, M., Takeshima, N., Hayasaka, Y., Imai, H., Shinohara, K., Tajika, A., Ioannidis, J. P. A., & Geddes, J. R. (2018). Comparative efficacy and acceptability of 21 antidepressant drugs for the acute treatment of adults with major depressive disorder: A systematic review and network meta-analysis. The Lancet, 391(10128), 1357–1366. https://doi.org/10.1016/S0140-6736(17)32802-7

Howick, J. (2011). Exposing the vanities—and a qualified defense—of mechanistic reasoning in health care decision making. Philosophy of Science, 78(5), 926–940. https://doi.org/10.1086/662561

Jauhar, S., Arnone, D., Baldwin, D. S., Bloomfield, M., Browning, M., Cleare, A. J., Corlett, P., Deakin, J. F. W., Erritzoe, D., Fu, C., Fusar-Poli, P., Godlewska, B., Hall, J., Harmer, C., Harrison, N. A., Harrison, P. J., Howard, R., Howes, O., Juruena, M. F., … Young, A. H. (2023). A leaky umbrella has little value: Evidence clearly indicates the serotonin system is implicated in depression. Molecular Psychiatry, 28, 3149–3152. https://doi.org/10.1038/s41380-023-02095-y

Kazdin, A. E. (2007). Mediators and mechanisms of change in psychotherapy research. Annual Review of Clinical Psychology, 3, 1–27. https://doi.org/10.1146/annurev.clinpsy.3.022806.091432

Moncrieff, J., Cooper, R. E., Stockmann, T., Amendola, S., Hengartner, M. P., & Horowitz, M. A. (2022). The serotonin theory of depression: A systematic umbrella review of the evidence. Molecular Psychiatry, 28, 3243–3256. https://doi.org/10.1038/s41380-022-01661-0

Sakrajda, K., & Rybakowski, J. K. (2025). The mechanisms of lithium action: The old and new findings. Pharmaceuticals, 18(4), 467. https://doi.org/10.3390/ph18040467

Son, Y. (2010). Molecular mechanisms of general anesthesia. Korean Journal of Anesthesiology, 59(1), 3–8. https://doi.org/10.4097/kjae.2010.59.1.3

Vane, J. R. (1971). Inhibition of prostaglandin synthesis as a mechanism of action for aspirin-like drugs. Nature New Biology, 231, 232–235. https://doi.org/10.1038/newbio231232a0

Vane, J. R., & Botting, R. M. (2003). The mechanism of action of aspirin. Thrombosis Research, 110(5–6), 255–258. https://doi.org/10.1016/S0049-3848(03)00379-7