You did not choose the nervous system you were born with. You did not choose the family system you were thrown into, the attachment patterns that formed before you had language for them, the early environment that taught your subcortical self whether the world could be trusted, or the early threats and uncertainties that calibrated your threat-detection threshold before your prefrontal cortex was developed enough to contextualize them. Heidegger called this Geworfenheit - thrownness - the condition of finding yourself already situated in a world you did not design, with a history you did not author, running on a nervous system that was shaped by all of it - cortisol reactivity, attachment insecurity, default threat-detection threshold, the capacity for self-regulation under stress - these were being built in you long before you had any say in the matter. Chronic anxiety, in many cases, is what that inheritance looks like when the threat-detection system has stayed dialed up past the conditions that originally taught it to do so.
Underneath that, woven into the architecture from the beginning, is biology.
- The genetic variants that reduce GABA synthesis rate and amplify the system's sensitivity to glutamate.
- The serotonin transporter polymorphisms that bias emotional reactivity toward the catastrophic interpretation of ambiguous input.
- The slow-clearance COMT variant that floods the prefrontal cortex with catecholamines under stress, producing the ruminative loop that won't quiet.
- The HPA axis that was sensitized early and has never fully recalibrated, keeping cortisol elevated when no present threat is there to justify it.
None of this biological soup - the specific systems that are underperforming, the genetic architecture that predisposed you, the inflammatory and hormonal variables that are actively driving the picture - was likely examined when you were handed a diagnosis, because seeing it fully requires the precise things that most clinical encounters aren't built to provide: time, compassion, and presence.
This guide is about the fact that the mechanisms herein are addressable - and to individuals already biased to think their nervous system is just the way it is, allow this small hope to be evidence to the contrary.
Before you begin
Is this for you?
This guide is for
- People with chronic or persistent anxiousness - the background hum of unease, physiological hyperarousal, difficulty unwinding, or a nervous system running at a higher gear than circumstances warrant
- Anyone whose situational anxiety has become generalized - the system has lost the ability to return to baseline between stressors, and the threat-response is now the default state rather than the exception
- The exhausted-but-wired phenotype - tired, unable to switch off, sleeping poorly, running on cortisol, the arousal no longer tied to any specific trigger
- People who have tried SSRIs or benzodiazepines and found them partially effective or blunting, or who want to understand what their medication is and is not doing at a mechanistic level
- Anyone with a mixed anxious-depressive presentation, or who cycles between hyperarousal and motivational shutdown - reading this in parallel with its sibling, A Field Guide to Depression & Low Mood, is typically more useful than either alone
This guide is not for
- Panic disorder with frequent, debilitating panic attacks - this addresses the chronic anxious baseline, not acute panic; significant functional impairment warrants clinical evaluation first
- OCD, PTSD, or phobia-specific presentations - these have distinct neurobiological profiles and treatment considerations a general anxiousness guide cannot responsibly address in full
What it is
FEAR through the time-travel lens
Anxiety is more than a neurochemical imbalance, more than a character pattern, more than a psychological problem - though it acquires psychological dimensions quickly once it has been present long enough to organize itself into avoidance patterns, catastrophic thinking, and identity. Each level of analysis catches a real facet of what's happening - the conventional models aren't wrong, they're working at different levels of evaluation. Before any of them, anxiousness is a primary affective circuit being recruited by something it wasn't built to respond to.
Affective Neuroscience and the Anxious Brain
The circuit is FEAR. Jaak Panksepp identified seven primary affective systems shared across mammals, with the underlying circuitry tracing back through vertebrate evolution into crustaceans and likely further still. Affect here means the conscious felt experience of an emotional or motivational state, the way an organism becomes aware that its homeostatic preferred states are deviating from baseline. Each system is anchored in distinct neural circuits with conserved anatomy, characteristic behaviors, and specific neurochemistries: SEEKING, RAGE, FEAR, LUST, CARE, PANIC/GRIEF, and PLAY. FEAR is the simple affective state whose imperative is to motivate the organism to avoid predators or move out of the instigating environment. The descending circuit runs through the dorsal periaqueductal gray (dPAG), the medial hypothalamus, the dorsomedial thalamus, and the centromedial amygdala via the amygdalofugal pathway. It evolved to do one specific thing: detect environmental danger and mobilize the response. Heart rate up. Respiratory rate up. Peripheral vasoconstriction. Freezing, fleeing, or - with enough RAGE recruitment - fighting. Action, fast.
The classical psychiatric interpretation that the amygdala "generates fear" is too simple. The amygdala plays multiple distinct roles, and conflating them is the source of much confusion. The basolateral amygdala (BLA) is the fear-learning module, building context-specific associations with previously encountered threats - a learning process that depends on intact hippocampal circuitry for the context-binding. The centromedial amygdala is one node in the descending FEAR output above. And the amygdala's forwarding of contextual fear information into the anterior cingulate cortex (ACC) and orbitomedial prefrontal cortex (omPFC) is what generates anxiety proper - the cognitive-contextual processing that runs on top of the unconditioned FEAR circuit, evaluated against memory and projected onto the future.
The descending FEAR circuit is one node of the broader mesencephalic decision-making triangle - the integrative machinery where the superior colliculus (target selection across spatial senses), the hypothalamus (motivation and need-state), and the periaqueductal gray (action selection and behavioral mobilization), together with input from the substantia nigra, perform the moment-to-moment computation of what to do next. Across vertebrate evolution, this is the integrative engine through which the entire forebrain - including the cortex - has to push its information when behavior (including emotive) is being selected. Prediction runs at every level of the neuraxis. The cortex elaborates; the mesencephalic triangle decides and acts. This is also why subcortical interventions (breathwork, somatic practice, vagal stimulation, voluntary cold exposure) reach a depth that purely cortical interventions cannot - they're operating on the integrative hub where the actual computation of what to attend to and what to do is happening.
That circuit is conserved across hundreds of millions of years of vertebrate evolution because it works. The animal that detects the predator and runs survives. The animal whose FEAR circuit was sluggish or absent does not. We are descended from organisms whose FEAR systems fired reliably in the presence of immediate threat.
Then, very recently in evolutionary terms, something happened that changed how this circuit gets used: autonoetic consciousness is the capacity for self-aware mental time-travel - the ability to construct a self that extends backward into a remembered past and forward into an imagined future, and to run the predictive model not just over present sensory states but over temporal sequences - over narratives. Plants, bacteria, and animals have varying degrees of consciousness, but autonoetic consciousness - the experience of being a self in time, reflecting on its own continuity - appears to be a primarily human elaboration.
The capacity is extraordinary, but it's also the source of the specifically human form of suffering. An animal in the FEAR circuit is afraid of something present. A human in the FEAR circuit can be afraid of something that happened three years ago and something that might happen next month simultaneously - neither of which exists as a current sensory state, both of which generate physiological arousal that the nervous system treats as if the threat were happening right now. Because at the level of the circuit, it is. At the level of the descending FEAR circuit, the distinction between a remembered tiger, an imagined one, and a real one doesn't exist - the dPAG mobilizes the same defensive output whether the threat signal arrives from present sensory evidence, the BLA-hippocampus learning module reactivating a memory, or the cortical generative model running an imagined future. The amygdala's forwarding to ACC/omPFC produces anxiety proper in all three cases. Responses are similar, although contextually variable, to all three.
Anxiety is FEAR applied to a future that hasn't arrived. Rumination is FEAR (and PANIC/GRIEF, the separation-distress circuit, which often runs alongside) applied to a past that can no longer be acted on. Both are second-order phenomena: legitimate first-order physiological arousal driven by simulated (either imagined or remembered) rather than present input. The body's response is identical in kind to a present-threat response, but the threat itself is a model running in the prefrontal cortex.
The prefrontal cortex - the most recently evolved structure in the human brain, the seat of planning, abstraction, and narrative self-construction - is both a gift and a liability. It extends the temporal horizon of the predictive model far beyond any other animal's. And it can generate suffering at that extended horizon without any present-moment threat to justify the cost.
Looking at anxiety from its true biological affective roots, and its neuroanatomical genesis, as opposed to through a purely neurochemical lens, changes what any intervention should aim at. The conventional psychiatric model treats anxiety as a malfunction in serotonin or GABA signaling to be pharmacologically corrected. The conventional psychological model looks deeper - at the events and experiences that shaped the nervous system, using cognitive and somatic reframing to update the felt relationship to them. With an astute therapist or coach, the psychological work is often extremely effective on its own. The mechanistic frame in this guide treats anxiety as a primary affective circuit (FEAR) being chronically recruited by a generative model decoupled from present sensory evidence, which provides the deepest available view of how the system was trained, why it keeps firing, and where the leverage actually sits.
Active Inference and the Anxious Mind
To understand the level the psychiatric and psychological models are both operating on, the framework that maps it most precisely is Karl Friston's free energy principle and the active inference architecture that follows from it.
The free energy principle is not a theory of brains. It's a theory of any system that persists. Friston's framework applies to a bacterium navigating a glucose gradient with the same logical force it applies to a human navigating a social situation - because the fundamental problem is identical in both cases. The system has a model of what its environment should look like, it samples the environment stochastically, and it acts - physically or biochemically - in the direction that reduces the discrepancy between model and reality. Minimize surprise. Maintain coherence. Persist.
Active inference is what this looks like in practice. The organism doesn't passively wait, it generates predictions about what its sensory states should be, evaluates the environment against those predictions, and acts on the environment to bring sensory reality into alignment with what was predicted. This is a very important basis for understanding the human brain: it is not a stimulus-response machine. It is a prediction machine that attempts to minimize variational free energy.
So the brain is a prediction machine that acts on the world to minimize variational free energy - the gaps between its predictions and reality. A prediction error in this framework isn't a malfunction or a sign that something is wrong. It's any discrepancy between what the model predicted and what arrived - some of which the system actively scans for, others it tries to avoid. Attention is generated by the precision-weighting of those errors - the organism's continuous decision about which discrepancies are worth resolving and which can be safely ignored. Think about what happens to your hearing when someone across a loud restaurant says your name. The acoustic environment hasn't changed. The signal-to-noise ratio of your name against the background noise hasn't objectively improved. But your auditory system has been running a low-priority scan for that specific pattern the whole time, and the moment it detects a match, it dramatically increases the precision - the weight, the trust - assigned to that prediction error and pulls your attention toward it against everything else. That's your nervous system selectively amplifying the prediction errors it has learned are most likely to matter. Attention isn't a spotlight you consciously aim. It's a precision dial your brain turns up on the signals it has decided, based on history and current state, are worth the cost of resolving.
There's a step in this story that pure information theory doesn't capture: how prediction error becomes a felt experience. Feelings are not byproducts of the prediction-error machinery - they are how prediction error becomes conscious, and being conscious is what allows the organism to act on it. Affect is intrinsically conscious; there are no unconscious feelings. A prediction error in pure information-theoretic terms is just a number. For the organism to actually adjust behavior, the discrepancy has to be felt. Anxiety, then, is the felt awareness of a prediction error the system has flagged as worth resolving - the conscious manifestation of an unattended-to need to act. The reason it's so gripping is that the felt-needs framework is a pre-cognitive priority system: feelings move behavior before conscious narrative catches up. The cortex eventually weaves a story about why the anxiety is there; by the time the story arrives, the body has already been mobilized.
Now apply this lens to chronic anxiety.
Rumination is active inference applied to the past, trying to resolve prediction errors that can no longer be corrected by action. The conversation that went wrong, the decision you regret, the relational rupture you keep replaying - the system is generating prediction errors against a model of how that situation should have gone, and it can't act on the past to reduce the error. So the error keeps firing. The replay isn't pathological in some abstract sense. It's the predictive machinery doing what it's built to do, in a domain where doing it is structurally incapable of producing resolution.
Anxiety is active inference applied to a future that hasn't arrived. The presentation next month, the medical result you're waiting on, the relational outcome you can't control. The system generates predictions about how that future will go, the predictions are weighted toward threat (because of past experience, genetic disposition, current stress load, or all three), and the system experiences those threat-weighted predictions as if they were present sensory data. Then, in an attempt to satisfy the predictions, the nervous system enacts its own evidence as physiological arousal, but the threat is only a model.
Chronic stress, trauma, and persistent threat states don't just make organisms feel bad. They recalibrate the precision dial. A nervous system that has learned, through repeated experience, that certain categories of signal reliably predict threat will assign high precision to those signals even in environments where the threat is absent. The hypervigilance of a trauma survivor isn't irrational - it's a precision-weighting problem. The model is accurate to the environment it was trained in. It's running that environment's threat-detection settings in a different environment, and the cost is a nervous system that can't afford to stop paying attention to things that no longer require it.
This is why anxiety so often persists past the circumstances that originally produced it. The model doesn't update easily. Once a precision dial has been turned up on a category of threat signal, ordinary positive experience tends to be weighted as exception or noise rather than as evidence the model should change. The system filters incoming evidence to confirm what it already expects, and in the absence of that confirmatory evidence will generate its own as the physiologic symptoms of anxiety. Updating it requires intervention specifically aimed at the precision-weighting itself - which is the mechanism behind why exposure therapy, EMDR, Somatic Experiencing™, Internal Family Systems™ work, certain meditative practices, and (where indicated) targeted neurostimulation actually work. They aren't suppressing the symptom. They're providing the kind of evidence the nervous system can no longer collect on its own because its filter has stopped letting that evidence through. For the trauma-survivor phenotype specifically - particularly where early-life adversity is part of the picture - the somatic and parts-based modalities (EMDR, Somatic Experiencing™, IFS™) often have stronger evidence than exposure therapy alone, because the precision-weighting was installed at a developmental stage that pre-verbal and pre-cognitive interventions reach more directly than top-down reframing does.
The combination of these two framings - FEAR circuitry recruited by mental time-travel, and active inference recalibrated by chronic threat - gives us the conceptual map. The biological hardware that executes all of it is what the rest of this guide addresses: the HPA axis, the autonomic pendulum, the hormonal system, the inflammatory milieu, and sleep architecture. Those are the biological hardware on which the prediction machine runs. They are also where most of the addressable interventions live.
The genetic architecture
Your brain's wiring
There is a genetic component that, in interaction with experience, configures how anxiously a given nervous system runs. Knowing your specific architecture - what your brain is wired toward - does not predict your fate, but it does explain enough to make the suffering make sense, and it does inform which interventions are most likely to help.
What follows is a clinically focused summary of the genetic variants most consistently associated with anxiety, rumination, stress reactivity, and impaired fear extinction. A detailed reference for each variant with allele effects and sources is in Appendix A. Most people reading this guide carry several of these variants, but two people carrying the same exact variants still won't predictably experience anxiety the same way or to the same degree.
The Serotonin System
Serotonin regulates the brain's capacity to tolerate uncertainty, recover from negative experiences, and maintain emotional equilibrium. Variants here are central to anxiety phenotype.
The serotonin transporter (SERT) determines synaptic serotonin clearance and is among the most extensively studied genetic factors in anxiety research. People with reduced transporter expression show, on average, greater emotional reactivity - higher amygdala activation in response to negative emotional stimuli (faces, threatening images, social rejection cues), greater sensitivity to stress and social context, and higher rates of anxiety disorders. People with higher transporter expression have a more emotionally stable baseline, but less responsiveness to SSRIs because the transporter is the SSRI target. Neither configuration is pathological in isolation; both are clinically informative. The reduced-transporter phenotype in a high-stress environment is the anxiety presentation that conventional medicine tends to encounter most often.
The 5-HT1A autoreceptor inhibits serotonin release when activated - it is the system's own brake. Variants that produce higher autoreceptor expression result in stronger self-inhibition, lower net serotonin output, and significantly higher rates of comorbid anxiety-depression. This is one of the variants most consistently associated with treatment-resistant presentations.
The Catecholamine System
Where serotonin variants shape emotional reactivity, catecholamine variants shape cognitive style and stress response in ways that map directly onto the rumination phenotype.
COMT (catechol-O-methyltransferase) is the most clinically actionable variant in the catecholamine system. The enzyme breaks down dopamine and norepinephrine in the prefrontal cortex, and there are two functional variants that meaningfully diverge in clinical presentation. The slow-clearance variant produces higher baseline dopamine and norepinephrine in the PFC, sharper cognition under normal conditions, and significantly worse performance under stress - this maps onto the ruminative anxiety pattern most cleanly: cognitive overactivation, catastrophic thinking loops, difficulty letting go of mental content, anxiety that lives in the head more than in the body. The fast-clearance variant produces lower baseline catecholamines and tends more toward motivational flatness and anhedonic depression than toward anxious rumination - though carriers can still develop anxiety from other genetic and environmental contributors. Knowing your COMT genotype meaningfully informs whether the supplement and lifestyle approach should emphasize calming the overactive (slow-clearance) system or supporting the under-recruited (fast-clearance) one.
Monoamine oxidase A (MAO-A) determines the rate of serotonin, norepinephrine, and dopamine breakdown after release. Variants that reduce enzyme activity produce slower monoamine breakdown, higher synaptic monoamine concentrations, and increased emotional reactivity. The classical "warrior gene" framing in the popular press oversimplified this: low-activity carriers in adverse early environments show increased risk for impulsive aggression and emotional dysregulation; the same carriers in stable environments often show no behavioral phenotype. This is one of the cleaner gene-by-environment interactions in psychiatric genetics, and it is directly relevant to the anxiety phenotype that emerged from a high-stress childhood.
The HPA Axis and Stress Response
These variants do not cause anxiety directly. They lower the threshold at which stress exposure translates into lasting neurobiological change - the molecular mechanism by which chronic stress recalibrates the precision dial described earlier.
FKBP5 regulates glucocorticoid receptor sensitivity and HPA axis recovery speed. Risk variants produce a system that takes longer to turn off after a stress response - chronically elevated cortisol exposure that promotes amygdala sensitization and sustains the threat-detection settings even after the threat has passed. The FKBP5 risk variant is the most consistently replicated gene-by-environment interaction in stress-related disorders. The variant alone doesn't predict anxiety; the variant in the context of early-life adversity or chronic adult stress substantially increases risk. This is the "the model is accurate to the environment it was trained in" architecture written in the genome.
NR3C1 is the glucocorticoid receptor gene itself. Risk variants reduce receptor transcriptional activity and increase glucocorticoid resistance - the brain is less responsive to cortisol's negative feedback signal, impairing HPA axis termination. Frequently co-implicated with FKBP5 in chronic anxiety presentations, particularly the kind that emerged from sustained early-life stress.
CRHR1 encodes the corticotropin releasing hormone receptor 1, which mediates the brain's initial threat response to CRH. Variants here affect receptor binding affinity and signal transduction efficiency - relevant to both the amplitude of the stress response and the anxious arousal that accompanies it. CRHR1 interacts with FKBP5 and NR3C1 to determine the overall stress-reactivity profile.
The GABAergic and Adenosine Systems
This cluster contains the variants most directly relevant to the GABAergic insufficiency and overactive arousal patterns described in the mechanism section.
GAD1 encodes glutamic acid decarboxylase 1 - the enzyme that converts glutamate to GABA. Risk variants reduce GAD1 expression or activity, impairing GABA synthesis and producing a state of relative glutamate excess and reduced inhibitory tone. This is a direct genetic mechanism for the GABAergic insufficiency that drives much of the felt experience of anxiety. Carriers often respond well to interventions that address GABA synthesis upstream (B6 as P5P, magnesium) rather than only modulating receptor activity.
ADORA2A encodes the adenosine A2A receptor. Caffeine acts as an adenosine antagonist; the A2A receptor is one of its primary targets. Risk-variant carriers experience markedly increased anxiety, panic, and cardiovascular reactivity in response to caffeine compared to non-carriers. If you have a clinical picture of generalized anxiousness combined with disproportionate adverse response to caffeine - panic, palpitations, jittery agitation lasting hours - this variant is probably part of your picture. The clinical advice for risk-variant carriers is straightforward: caffeine is not a default-acceptable input for you, and the question is not whether to drink it casually but whether the cost-benefit ever justifies it for your specific neurochemistry.
Neuroplasticity and Fear Extinction
Variation in BDNF (brain-derived neurotrophic factor) is the single most relevant genetic factor for why some anxious presentations respond well to therapy and exposure-based interventions and others don't. BDNF supports synaptic plasticity and learning, and is centrally involved in fear extinction - the brain's capacity to learn that a previously feared stimulus is now safe. The risk variant produces reduced activity-dependent BDNF secretion, impairing fear extinction at the molecular level. Carriers often show more persistent anxiety responses, slower extinction of conditioned fear, and weaker response to exposure-based interventions. Clinically, this means a longer therapeutic timeline, additional pharmacological support during extinction work, or combined interventions (such as exercise, which upregulates BDNF) to produce durable change.
What To Do With This
You do not need genetic testing before beginning this protocol. The approach was designed with the full range of these phenotypes in mind. What genetic information does is help you weight the protocol toward your specific picture - the slow-clearance COMT ruminator centers magnesium, L-theanine, and downregulation of cognitive arousal; the FKBP5 carrier with early-life stress centers ashwagandha and HPA recalibration alongside the trauma-informed work that addresses the precision dial directly; the BDNF risk-variant carrier benefits from exercise as a non-negotiable adjunct rather than an optional addition.
If you want genetic information, you have a few practical paths. Direct-to-consumer raw data testing through 23andMe (~$100, results in 4-6 weeks) gives you the raw SNP file. That file then needs interpretation - the cleanest tools for the variants in this guide are Strategene by Genova Diagnostics (automated interpretation focused on clinical relevance, you can upload your existing 23andMe data), Genomind (clinical-grade, requires prescriber order), or GeneSight (psychiatric-medication focused, prescriber-ordered). If you want a clinical interpretation alongside the report, work with a functional-medicine practitioner who reads these panels regularly. Genetic testing is informative, not prescriptive - your specific protocol should still be guided by your clinical phenotype, lab markers, and response to interventions. See Appendix A for the full variant reference.
The mechanism
What's happening in the body
The conceptual frame above describes what anxiety is computationally and affectively. The biological substrate that executes it - where supplements, lifestyle interventions, and (where indicated) medications actually act - is the integrated activity of the HPA axis, the autonomic pendulum, the hormonal system, the inflammatory milieu, and sleep architecture. Which of those is most dysregulated in your case determines which intervention is most likely to work, and which produces only temporary symptomatic relief while the underlying mechanism keeps running.
The HPA Axis
The hypothalamic-pituitary-adrenal axis is the central stress response system. Under normal conditions, it operates phasically - it fires in response to a threat, cortisol rises, the organism responds, the threat resolves, cortisol returns to baseline. The system is designed for bursts, not for continuous activation.
Chronic psychological stress, poor sleep, caloric restriction, gut dysbiosis, and systemic inflammation all constitute low-grade, unresolvable threats that keep the HPA axis in a state of tonic activation. Cortisol does not return to baseline because the threat signal never clears. Over time, this produces a dysregulated cortisol curve - often elevated in the evening when it should be at its nadir, and blunted in the morning when the cortisol awakening response should be sharp. The result is the exhausted-but-wired state: physiologically aroused, experientially depleted, unable to switch off.
CRH - corticotropin releasing hormone, the hypothalamic signal that initiates the HPA cascade - is itself anxiogenic. This is the mechanism by which the stress response and the subjective experience of anxiety are coupled at the molecular level. It is not that you feel anxious because you are stressed. It is that the same molecule that activates the stress response directly activates the neural circuits that generate the phenomenology of anxiety. The throttle and the feeling are connected at the source.
The Autonomic Pendulum
The autonomic nervous system is the pendulum that swings between activation and recovery. The sympathetic branch drives the threat response: heart rate up, respiratory rate up, peripheral vasoconstriction, muscle tension, digestion suspended. The parasympathetic branch - primarily through the vagus nerve - drives recovery: heart rate down, respiratory rate down, digestion restored, immune function supported. The healthy nervous system swings cleanly between the two states; the anxious nervous system is weighted toward the activated side and has lost the capacity to swing back.
Chronic anxiousness involves chronically elevated sympathetic tone and, critically, a reduction in parasympathetic tone - the capacity of the parasympathetic system to actively return the organism to rest. Parasympathetic tone is measurable as heart rate variability (HRV): the beat-to-beat variation in heart rate that reflects the continuous push-pull between sympathetic and parasympathetic influence. Low HRV is both a marker and a consequence of chronic anxiousness, and it is independently associated with cardiovascular risk, cognitive impairment, and poor stress resilience.
The vagus nerve is bidirectional - it carries signals from the brain to the body and from the body to the brain. This is why slow, diaphragmatic breathing reliably reduces acute anxiety: it directly stimulates vagal afferent fibers that signal safety to the brainstem, dampening the descending FEAR-circuit response via the nucleus tractus solitarius. The breathing changes the hardware state, which changes the phenomenological state.
GABAergic neurochemistry mediates the brake side of the pendulum - it allows the parasympathetic system to actually dampen sympathetic tone. GABA (gamma-aminobutyric acid) is the primary inhibitory neurotransmitter in the central nervous system. Its job is to reduce neuronal excitability: to apply the brake to circuits that are firing too hard or too persistently. Anxiousness, at the neurochemical level, is in large part a state of insufficient GABAergic tone across multiple levels - cortical (rumination, catastrophic thinking), limbic (amygdala hyperreactivity), midbrain (PAG-driven defensive mobilization), and brainstem (sympathetic disinhibition, reduced vagal output). The dampening of sympathetic tone and the active engagement of parasympathetic recovery both depend on GABAergic signaling at the brainstem level, particularly in the nucleus tractus solitarius and the rostral ventrolateral medulla.
I've always recommended thinking about benzodiazepines as break-glass-in-case-of-emergency tools, not as a stash to keep on hand for managing anxiety. The mechanism is reliable and immediate: they're positive allosteric modulators of the GABA-A receptor, amplifying the response to GABA at every synapse where it is active - across cortex, limbic system, and brainstem. What they don't do is touch the prediction error driving the arousal in the first place. The model that generated the threat-weighted forecast is still running; the body's enactment of that forecast - the physiological arousal - just got chemically suppressed. Symptom relief without model update - which is what makes benzos a chemical bypass. Useful in the right moment. Costly as a default tool. The consequences of chronic receptor modulation at that breadth - tolerance, dependence, suppression of slow-wave sleep, and the rebound hyperexcitability that makes discontinuation difficult - are well-documented and predictable.
GABAergic insufficiency can arise from multiple upstream causes: magnesium deficiency (magnesium is a physiological modulator of GABA-A receptor sensitivity), B6 deficiency (pyridoxal-5-phosphate is a required cofactor for the enzyme that synthesizes GABA from glutamate), chronic alcohol use and withdrawal, and the glutamate excess that accompanies chronic stress and poor sleep. The intervention does not always require targeting the receptor directly - sometimes it requires fixing what is upstream of GABA synthesis.
The autonomic system at the level of organs and the GABAergic system at the level of synapses aren't parallel - they're the same pendulum described at different levels. The pendulum's swing depends on the brake working. The brake depends on adequate GABAergic tone. When GABAergic tone is insufficient, the pendulum gets stuck on the activated side - which is what anxiety looks like in the body.
The Hormonal Dimension
Anxiety has a hormonal axis worth understanding in its own right. For roughly half the population, the dominant timing variable in their anxiety pattern is hormonal, not neurochemical alone. Several of the body's most powerful endogenous anxiolytics and anxiogenics are sex steroids and thyroid hormones, and their fluctuations across the menstrual cycle, across seasons, and across the lifespan reorganize the entire HPA-autonomic pendulum picture described above.
Progesterone is the primary endogenous source of positive allosteric modulation at the GABA-A receptor. Specifically, its neuroactive metabolite allopregnanolone is a potent positive allosteric modulator of the GABA-A receptor at the neurosteroid binding site, distinct from but functionally analogous to the benzodiazepine binding site. When progesterone is high, GABAergic tone is amplified and the nervous system runs at a lower baseline arousal - this is why many women feel calmer in the mid-luteal phase when progesterone peaks, and why the pre-menstrual progesterone withdrawal produces a distinctive anxiety phenotype that maps onto sudden GABA-A under-modulation. Premenstrual dysphoric disorder (PMDD) is the severe end of this pattern - a four-to-fourteen-day window of marked anxiety, irritability, dysphoria, and physical symptoms that resolves at the onset of menstrual flow. PMDD is not a mood disorder that happens to be cyclical. It is a cyclical neurosteroid withdrawal phenomenon with anxiety-disorder phenomenology.
Estrogen modulates serotonin synthesis and 5-HT1A receptor density. When estrogen is high, serotonergic tone is supported. When it drops - across the menstrual cycle, in perimenopause, postpartum - serotonin-dependent emotional regulation can destabilize. This is part of why many women experience anxiety symptoms in the late luteal phase (estrogen drops alongside progesterone), perimenopause (estrogen oscillates erratically before declining), and postpartum (estrogen and progesterone both fall by orders of magnitude in 24-48 hours after delivery). Postpartum anxiety, distinct from postpartum depression, is increasingly recognized as a presentation that often goes undiagnosed because the postpartum mental health screening is depression-focused. The recent FDA-approved neurosteroid analogs (brexanolone IV, zuranolone oral) are approved specifically for postpartum depression - the Depression guide's hormonal section covers the mechanism in full - but they target the same allopregnanolone-withdrawal biology that drives postpartum anxiety, and the shared etiology means these treatments are mechanistically relevant here even where the formal indication is not.
Perimenopause is the multi-year endocrine transition where the cyclical neurosteroid scaffolding that has supported GABAergic tone for decades begins to break down. The clinical picture is often dramatic - anxiety, panic, sleep disruption, mood volatility, cognitive changes, and a felt sense of "being in a stranger's body" - and is frequently misattributed to psychiatric primary cause when the driver is endocrine. This is one of the largest single drivers of new-onset anxiety in women aged 40-55, and the conventional response (SSRI, sometimes benzodiazepine) addresses the downstream signaling without touching the upstream withdrawal. The full picture lives in A Field Guide to Perimenopause. For the anxiety presentation specifically, the upshot is that the supplement protocol below works much better when hormones are supported than against an unaddressed perimenopausal deficit.
Testosterone matters in both sexes. Low T contributes to increased anxiety, decreased stress resilience, and sleep disruption regardless of cause - and treating the anxiety without addressing the upstream picture produces partial response. The biological evaluation comes first. Chronically elevated cortisol suppresses testosterone production at the source, environmental endocrine disruptors interfere with steroidogenesis, sleep apnea cuts overnight production, gut and metabolic inflammation suppress the HPG axis, and a handful of less common causes (pituitary, primary testicular) round out the differential. Many men whose serum T comes back low present sedentary and withdrawn not by choice but because the biology is disrupted - addressing what's actually upstream is what restores the capacity, not the other way around.
In women, low testosterone (often dropping in perimenopause and postpartum) contributes to anhedonia and reduced stress resilience that overlaps with anxiety presentations.
The full clinical picture lives in A Field Guide to Low T.
Thyroid hormone is closely coupled to anxiety. Subclinical hyperthyroidism (suppressed TSH with high-normal or elevated fT3) produces tachycardia, heat intolerance, tremor, and physiological hyperarousal that maps directly onto generalized anxiety and is missed by TSH alone. Hypothyroidism with elevated reverse T3 (the inactive thyroid metabolite that builds up under chronic stress) produces the opposite picture but can coexist with anxious presentations through its impact on energy and resilience. It's worth running a full thyroid panel including fT3, fT4, and reverse T3 if the anxiety picture has any thyroid-suggestive features.
The relevant labs for the hormonal axis: progesterone (day 21 of a typical 28-day cycle, or a week after confirmed ovulation, for cycling women), estradiol, total and free testosterone with SHBG, DHEA-S, full thyroid panel (TSH, fT3, fT4, reverse T3, TPO and Tg antibodies), and cortisol curve as already noted.
Menstrual-cycle tracking (apps like Mira, Inito, Natural Cycles, or Flo) is informative regardless of intervention plan - the pattern of when your anxiety appears across the menstrual cycle tells you what the dominant hormonal driver is.
If your anxiety has any cyclical, postpartum, perimenopausal, or thyroid-suggestive component, the protocol below is not the first line - the hormonal evaluation is. The supplement stack works much better when hormones are supported than against an unaddressed hormonal deficit.
The Inflammatory Model
Inflammation isn't usually a contender in people's minds when it comes to anxiety; however, the absence of joint pain or body soreness doesn't mean inflammation isn't part of the picture. Neuroinflammation can be active without the somatic markers people typically associate with the inflammatory state. Peripheral inflammatory cytokines (IL-6, TNF-alpha, IL-1beta) cross the blood-brain barrier, activate microglia (the brain's resident immune cells), and produce a coordinated neuroinflammatory state that sensitizes the descending FEAR circuit, suppresses GABAergic tone, dysregulates the HPA axis, and shifts neurotransmitter synthesis pathways toward neurotoxic metabolites. An inflamed brain is a more anxious brain through direct molecular pharmacology happening simultaneously at multiple sites.
One pathway worth understanding specifically is the kynurenine shunt. Inflammatory cytokines upregulate the enzyme indoleamine 2,3-dioxygenase (IDO), which diverts tryptophan away from serotonin synthesis and toward the kynurenine pathway. The result is two simultaneous problems: less serotonin available for emotional regulation, and increased production of quinolinic acid - an NMDA receptor agonist that is directly excitotoxic and contributes to glutamate-driven anxiety states. This is why chronically inflamed individuals often have anxiety presentations that look serotonergically deficient on paper but respond poorly to SSRIs - the bottleneck isn't serotonin reuptake, it's serotonin synthesis being shunted away by an inflammatory signal that the SSRI doesn't touch.
Phenotypes that warrant explicit inflammatory framing:
Autoimmune-driven anxiety. Hashimoto's thyroiditis, lupus, rheumatoid arthritis, celiac disease, and inflammatory bowel disease all present with anxiety symptoms that conventional psychiatry treats as comorbid mental health conditions. They aren't coincidental comorbidities - they share the inflammatory milieu that drives the anxiety presentation, even when the autoimmune target tissue differs. Hashimoto's specifically deserves explicit mention: the antibody-mediated thyroid inflammation produces an anxiety phenotype even when TSH is normal, and the antibody titers (TPO, thyroglobulin) often correlate with symptom intensity better than thyroid hormone levels do.
Diet- and gut-driven anxiety. Approximately 80% of vagal fibers run from gut to brain rather than brain to gut - the gut is sending more information upward than the brain is sending down, and diet is the primary daily input shaping that signal. Ultra-processed food, high-glycemic carbohydrates, industrial seed oils, and (in genetically susceptible individuals) gluten and casein generate a cumulative inflammatory load through gut dysbiosis, intestinal permeability, and the resulting endotoxemia (lipopolysaccharide reaching the systemic circulation), which reaches the brain via vagal afferents and the systemic circulation simultaneously. Insulin resistance compounds this through adipose-derived inflammatory cytokines and is independently associated with anxiety risk. The Standard American Diet, in this frame, is an anxiogenic input. The anxious person with bloating, irregular transit, food reactivity, or post-meal mood disturbance is not coincidentally experiencing both - the gut state and the neural state are mechanistically coupled through the gut-brain-immune axis. This is why interventions like fasting or low-variability diets like carnivore can have dramatic effects on inflammatory mental health presentations - they remove the antigenic and microbial provocation keeping the inflammatory signal active, letting the system quiet without daily re-stimulation. Even still, if there's active dysbiosis, dietary input reduction starves the overgrowth but doesn't necessarily clear it - the antimicrobial, herbal, and biofilm-disruption work is its own clinical lift. See A Field Guide to Dysbiosis & SIBO for that protocol.
Post-viral and post-infectious anxiety. Long COVID, post-EBV syndromes, post-Lyme presentations, and post-mononucleosis fatigue states frequently include prominent anxiety features that emerged after the acute infection cleared. The mechanism is persistent neuroinflammation following viral insult - particularly with the herpes family (EBV, HSV-1/2, VZV, HHV-6) and other pathogens with nerve-tissue tropism. Microglial priming after the initial infection lowers the activation threshold for subsequent inflammatory signals, blood-brain barrier disruption from systemic cytokine surge lets peripheral inflammatory mediators reach central tissue more readily, and PAMPs and DAMPs (pathogen- and damage-associated molecular patterns) keep activating innate immune receptors like TLR3/4/9 long after the acute infection has cleared. The immune system is still running an inflammatory program against signals it has learned to treat as threats. If your anxiety dates clearly from a specific infectious event, the workup should include viral antibody panels (especially herpes family), inflammatory markers, and consideration of antiviral or immunomodulatory approaches alongside the anxiety protocol.
Sleep-driven inflammation. Insufficient sleep elevates IL-6 and TNF-alpha within 24-48 hours, creating a circular pattern: poor sleep → inflammation → FEAR-circuit sensitization → anxiety → impaired sleep. The sleep section below addresses this.
Useful inflammatory markers for an anxiety workup: hsCRP (general systemic inflammation - target <1.0 mg/L), ferritin (chronically elevated suggests inflammation), fibrinogen, homocysteine (methylation and inflammation marker), IL-6 and TNF-alpha (research-grade, expensive), and antibody panels for the autoimmune presentations described above. A standard CBC with differential can reveal chronic inflammation through neutrophil-to-lymphocyte ratio elevation. Comprehensive stool testing adds detail the serum panel can't see - calprotectin (intestinal inflammation), zonulin and occludin (tight junction integrity and intestinal permeability), floral composition and dysbiosis patterns (commensals, opportunists, pathobionts, yeast), and secretory IgA (mucosal immune tone). Each gives direct visibility into the gut-driven inflammatory activity that can be running silently while the standard blood labs read clean.
Microglial priming is the longer-term consequence. Chronic inflammatory signaling teaches microglia to maintain a hyper-reactive state - they release more cytokines in response to subsequent triggers, and the threshold for full-on neuroinflammatory cascades drops over time. This is part of what makes chronically inflamed anxiety so treatment-resistant: even after the initial inflammatory driver is addressed, the microglia continue to over-respond until the priming reverses, which takes months to years and benefits from interventions that specifically modulate microglial activation (omega-3, resolvins, vagal stimulation, ketogenic-direction nutrition, exercise).
Treating the Inflammatory Front
The interventions in this guide that target inflammation specifically: omega-3 (especially the SPM Supreme formulation that delivers pre-formed resolvins and protectins), zembrin via PDE4 inhibition, breathing-driven vagal activation (the cholinergic anti-inflammatory pathway is well-documented and substantial), and adequate sleep. Beyond this guide: gut work, food-sensitivity testing where indicated, and addressing the autoimmune drivers directly in collaboration with the appropriate clinician. Inflammation is not always the primary driver, but when it is, the rest of the protocol works dramatically better once it is addressed.
Sleep Architecture
Sleep is among the most modifiable variables affecting anxiety, and conventional psychiatric framing tends to treat it as a downstream symptom rather than recognize its fundamental upstream position. The relationship is bidirectional, of course, but the upstream origin tends to be forgotten when it comes to intervention - configuring the variables in someone's sleep environment is far more difficult than treating the 'problem' anxiety with a pill.
Insufficient REM impairs the fear-extinction machinery that downgrades emotional charge on fear-laden memories - without it, yesterday's anxious experience remains a high-precision threat-prediction in the generative model rather than getting attenuated into a memory of a thing that already happened (BDNF risk-variant carriers depend especially heavily on adequate REM for this process). Slow-wave sleep loss prevents the overnight glymphatic clearance of metabolic waste; chronic short sleep elevates IL-6 and TNF-alpha within 24-48 hours, feeding the inflammatory-cytokine load that sensitizes the descending FEAR circuit. The cortisol descent that should hit its nadir between 11pm and 3am fails when sleep is fragmented, producing the 2-4am awakening with the mind turning on immediately. A single night of four-to-five hours produces next-day amygdala hyperreactivity, prefrontal inhibitory failure, glutamate accumulation in the PFC (reducing the GABAergic brake), heightened dPAG-driven autonomic mobilization, and impaired fear extinction - most chronically anxious people are running some version of this cascade chronically without recognizing it as the proximal driver of their anxiety floor.
Each of these mechanisms ultimately converges on the same downstream effect: degraded prefrontal regulation of the descending FEAR circuit and the amygdalo-cortical forwarding that produces anxiety proper. Sleep loss is not a peripheral lifestyle variable - it is a direct route into the same affective-circuit machinery this guide has been addressing, entering it from a different angle. The actionable target is 7-9 hours, with sufficient slow-wave (typically 13-23% of total, weighted toward the first half of the night) and sufficient REM (typically 20-25%, weighted toward the second). If your anxiety has any sleep-disruption component, A Field Guide to Sleep Architecture carries the full mechanistic picture and runs in parallel with this one.
The audit
Somatic awareness
Before any intervention, characterize your specific pattern. Anxiousness is not one thing - it's an underlying state expressing through different combinations of drivers, and the dominant ones determine the entry point. Most people have more than one driver active; the audit is about recognizing which ones rather than picking only one.
Temporal direction. Where is your anxiety pointed? Anxiety pointed at the future (anticipation, dread, "what if") is FEAR applied to mental time-travel forward. Anxiety pointed at the past (rumination, replay, "I can't stop thinking about") is FEAR plus PANIC/GRIEF (the separation-distress circuit) running on memory. Anxiety pointed at the present (moment-to-moment scanning, hypervigilance) is the precision dial stuck on threat without a clear temporal target. Most people have a dominant direction. The direction shapes which interventions land - structured worry-time for future-pointed, present-moment practice for past-pointed, breathwork and somatic grounding for present-scanning.
Cognitive vs. physical predominance. Is your primary experience physical - muscle tension, chest tightness, racing heart, shallow breathing, GI disturbance - or cognitive - rumination, worry loops, catastrophizing, difficulty concentrating? Physical predominance points toward autonomic and HPA mechanisms. Cognitive predominance points toward GABAergic insufficiency and impaired prefrontal regulation of the FEAR circuit, often the slow-clearance COMT ruminator profile.
Timing pattern. Is anxiousness present continuously, or does it follow a pattern? Morning anxiety that lifts through the day suggests low morning cortisol and HPA dysregulation (often a flattened cortisol awakening response). Evening anxiety and inability to wind down, or 2-4am awakening with the mind turning on immediately, suggests elevated nocturnal cortisol and sleep-cortisol coupling failure - the exhausted-but-wired phenotype. Anxiety that tracks your menstrual cycle, or that arrived with a clear postpartum or perimenopausal onset, implicates progesterone and estrogen withdrawal patterns - see the Hormonal Dimension section above.
Baseline vs. trigger. Can you identify consistent triggers, or has the anxious state become the baseline from which triggers pull you further? Triggers-against-a-resting-baseline is the earlier-stage anxious system. Anxious-baseline-from-which-triggers-pull is the HPA dysregulation pattern - the nervous system has lost the ability to return to a genuine resting state between events. The latter often means an FKBP5 / chronic-stress / early-adversity gene-by-environment pattern has consolidated.
Fear extinction. Is there a specific anxiety - a discrete fear, a phobia, a memory that hasn't downgraded, a relational pattern - that you "know is irrational but can't shake"? The "knowing it's irrational" is your prefrontal cortex correctly modeling that the threat isn't current. The "can't shake it" is the FEAR memory still holding precision-weighting in the generative model. The BDNF risk-variant fear-extinction-impaired pattern often presents this way and depends heavily on intact REM sleep for the molecular machinery of extinction to operate.
Sleep architecture. Anxiousness and sleep disruption are bidirectionally causal. Waking at 2-4am with a mind that turns on immediately is the nocturnal cortisol elevation pattern. Difficulty falling asleep despite exhaustion is sympathetic hyperarousal preventing the parasympathetic transition into sleep. Chronic short sleep without acute insomnia is the deprivation-cascade phenotype - running the cascade chronically without an acute symptom signal flagging it. A Field Guide to Sleep Architecture addresses the sleep side in more depth.
HRV and parasympathetic tone. If you track HRV with a wearable, chronic low HRV is the autonomic objective correlate. Trending HRV over weeks is more informative than single-night readings. Low HRV that doesn't recover with weekend rest implicates the sympathetic-dominance, low-parasympathetic-tone phenotype - the autonomic pendulum stuck on the sympathetic side.
Caffeine response. If caffeine produces disproportionate anxiety, panic, or persistent jitteriness, ADORA2A adenosine-receptor variation is probably part of your picture. If caffeine doesn't produce panic but you "can't turn it off" once you've had it - rumination intensifies, the mind churns longer than the stimulant effect should last - that's the slow-catecholamine-clearance pattern (slow COMT). Treat caffeine response as biological data, not personal failing.
Inflammatory and autoimmune picture. Persistent fatigue disproportionate to activity, gut symptoms (bloating, irregular transit, food reactivity, post-meal mood disturbance), joint pain, known autoimmune diagnosis, or anxiety that emerged after a specific infectious event - any of these suggest inflammation is part of your picture. The sub-phenotypes look meaningfully different: autoimmune (Hashimoto's especially - antibody titers often correlate with anxiety better than thyroid hormone levels), gut-and-diet driven (the dysbiosis-inflammation-vagal coupling), and post-viral (Long COVID, post-EBV, post-Lyme - mast cell activation and persistent low-grade neuroinflammation). See the Inflammatory Model section.
Hormonal driver (especially for women). Cyclical pattern (worsening in the luteal phase, lifting at menses), postpartum onset, perimenopausal age with sleep disruption and hot flashes, irregular cycles. Each implicates a different cluster within the broader hormonal axis. See the Hormonal Dimension section above and A Field Guide to Perimenopause for that specific transition.
Inheritance and early-life context. When did the anxiety start? Has it been present since childhood, or did it emerge later? Was there sustained early stress, adverse childhood experiences, or chronic adult stress before the consolidation? Genetic-level factors aren't deterministic, but the FKBP5 gene-by-environment finding is consistent: the genetic contribution plus an environment of sustained stress is what produces the chronically dysregulated phenotype. Recognizing the early-life input shapes whether trauma-informed work needs to be part of the entry point alongside the mechanism-level work.
If you have recent labs, the most useful markers for anxiousness specifically are: salivary cortisol curve (four-point diurnal, including a cortisol awakening response point), red blood cell (RBC) magnesium level (not serum - serum magnesium is tightly regulated and will be normal until deficiency is severe), hsCRP, vitamin D, ferritin, omega-3 index (RBC EPA + DHA as a percentage of total fatty acids; target above 8%, below 4% is independently associated with increased anxiety and inflammatory load), full thyroid panel (TSH, fT3, fT4, reverse T3, TPO and Tg antibodies), and a hormonal panel including progesterone (day 21 of a 28-day cycle, or a week after confirmed ovulation, for cycling women), estradiol, total and free testosterone, SHBG, and DHEA-S.
This is not an exhaustive list.
Before the bottles
Lifestyle scaffolding
The nervous system recalibration that resolves chronic anxiousness requires behavioral and physiological inputs that no capsule provides. The supplements in this guide are working on receptor sensitivity, neurotransmitter availability, and HPA axis tone. The lifestyle work below is working on the same systems at a level of magnitude that the supplements cannot match alone. Do both.
Breathing and Cold Exposure
Slow diaphragmatic breathing is the highest-ROI intervention in this guide relative to time invested and it costs nothing. The mechanism is specific: respiratory rate below 6 breaths per minute maximally stimulates the baroreceptor reflex and vagal afferent signaling, directly suppressing sympathetic tone and activating the parasympathetic system. The 4-7-8 pattern (inhale 4 counts, hold 7, exhale 8) is the cleanest example because the exhale phase is the active component - a longer exhale relative to the inhale is what drives parasympathetic activation. Five minutes twice daily produces measurable HRV improvement within two weeks of consistent practice.
Cold exposure is fundamentally an autonomic-pendulum exercise. The acute response is sympathetic, not parasympathetic - massive catecholamine release, peripheral vasoconstriction, gasping inhalation, heart rate spike. What makes the practice therapeutic is the recovery half of the pendulum: the parasympathetic rebound that arrives when the system is allowed to stand back down. Choosing to step into a stimulus that produces fear and physiological arousal, then regulating through it with conscious intent, is somatic-level practice for the precision dial. The system gets to learn, in real time, that fear and arousal can be tolerated and integrated rather than fled from. The catecholamine surge is what the practice operates on; the regulation during the surge is what makes the exposure therapeutic rather than just stressful.
If you can't regulate through the stimulus, it isn't therapeutic in this lens. White-knuckling, dissociating, performing for an audience, fighting through under peer pressure - none of that trains the precision dial; it trains the opposite. The system learns that fear and arousal are things to mask and push through under social observation, which is what generates anxiety, not what resolves it. The therapeutic window requires the capacity to remain present and self-regulating across the surge. Without that, you're just teaching the system to crank harder.
Movement
Movement reduces amygdala reactivity, upregulates GABA synthesis, and normalizes the HPA axis response to subsequent stressors - the trained individual mounts a smaller cortisol response to psychological stress and recovers faster. That much is well-established.
The deeper question - how training the body with intention generates interoceptive evidence that calibrates the predictive system itself - is developed in the Embodiment Anxiety essay and the Know Thyself Through Form essay. The short version: a well-mapped body is a more-predictable signal source for the predictive model, and that legibility lowers the interoceptive uncertainty that the threat-weighting system would otherwise mobilize against. Training with intention is foundational work for the anxiety phenotype, not just a state-change tool. For the full argument, those essays are the right place.
For the anxiety phenotype specifically, two practical points sit downstream of that broader frame.
First, movement used as coping versus movement used as foundational work are not the same intervention. The runner-for-stress-relief pattern is familiar: stress builds, you go for a run, you come back feeling regulated, you tell yourself the run handled it. The autonomic pendulum does swing, endorphin and endocannabinoid release produces acute relief, baseline cortisol comes down for a few hours afterward. The catch is found in the deeper architecture - you're literally running away from the stressors, the runner's high is masking the underlying load, and the system comes back regulated but not resolved. Tomorrow's anxiety floor is the same. The alternative that works deeper is training with intention - presence in the form, awareness of the breath-load coupling, attention to what's happening in the body during the work - which is what produces the calibration the Embodiment Anxiety essay develops.
Second, the format question. For the autonomic dimension specifically, short-window high-intensity work followed by complete recovery to baseline trains the autonomic pendulum to swing both ways. One to three minutes of effort that pushes physiology past clear thresholds (heart rate over 140-150, elevated systolic pressure, full perspiration, breathing rate rising into respiratory-compensation territory with CO2 buildup and bicarbonate loss), then complete rest until the system actually returns to baseline before the next set. The hard-work-then-full-recovery cadence teaches the system to swing both ways - the somatic version of what voluntary cold exposure does for the precision dial. Resistance training fits this format naturally and adds benefit through testosterone and BDNF mechanisms.
Two caveats. In individuals with significantly dysregulated HPA axes, prolonged endurance training adds extended cortisol elevation against an already overloaded system, and the acute relief masks cumulative damage being done underneath. Short-burst pendulum-style training is qualitatively different and generally well-tolerated even in dysregulated phenotypes - the cortisol stimulus is brief and the recovery half of the pendulum is built into the format. If you feel worse after a hard set rather than better, the issue may be inadequate recovery between sets rather than the intensity itself.
Timing also matters more in HPA-dysregulated phenotypes than in unaffected ones. Morning high-intensity work can be used deliberately to provoke the cortisol awakening response in flat-CAR phenotypes. Evening high-intensity training is more nuanced - the acute cortisol bump can be partially or fully offset by adenosine accumulation, glycogen depletion, and the parasympathetic rebound that builds substantial sleep pressure, but in some individuals shows up as disrupted sleep onset instead. The literature is mixed and individual response varies considerably. Track your own pattern.
Sleep
Chronic sleep deprivation is among the most reliable inducers of anxious arousal in otherwise healthy individuals, and the mechanism section above details why. If sleep is severely disrupted, address it in parallel with this protocol. See A Field Guide to Sleep Architecture. The two issues are mechanistically coupled and neither resolves fully without the other being addressed.
Caffeine
Caffeine blocks adenosine receptors and elevates catecholamines - both of which increase physiological arousal. In individuals with a genetically slow CYP1A2 metabolism, an ADORA2A risk allele, or already elevated baseline sympathetic tone, caffeine is an anxiogenic input that is worth auditing honestly. The question is not whether you feel jittery from caffeine - many anxious people are habituated and do not - but whether your baseline arousal state would be lower without it. A two-week caffeine elimination is one of the more informative self-experiments available to the chronically anxious person.
Digital Inputs
The phone is the most ubiquitous anxiogenic input in modern life, and it is structurally underweighted in clinical conversations. Four mechanisms: intermittent reinforcement of the SEEKING circuit through unpredictable notification rewards, which dysregulates the dopaminergic precision-weighting in ways that produce a chronic low-grade craving-driven arousal state; cortisol elevation from threat-content consumption, particularly news doomscrolling and conflict-laden social media, which the FEAR circuit reads as an unbroken stream of distant threats it cannot act on; sympathetic activation from the mere presence of the phone in your visual field - the anticipation of vibration produces measurable autonomic activation even when no notification arrives; and social comparison, which can feed either aspiration or envy depending on how it's metabolized. Aspiration accepts that the people on the feed got where they are through time, work, and struggle most of which is invisible, and uses the visible result as motivation to do the same - the predictive model updates to include "the journey, including the suffering it carries, is part of the cost." Envy is the failure mode: wanting the destination without the journey, which trains the model to anticipate progress curves no actual life produces. The chronic mismatch between the modeled life and the lived one registers as inadequacy, being-behind, defeat - prediction error in the direction of "I should be further along than I am" running continuously, recruiting threat-weighted attention to a discrepancy that has no available action. The thing to track when scrolling is which mode you're in.
A more recent twist is worth flagging: the influencer pattern of feigning vulnerability and mental-health distress for engagement. Vulnerability is an extraordinary quality, and one of the few things that builds genuine trust and intimacy between humans - which is precisely why weaponizing it for content is corrosive. When suffering becomes performance and "struggle" becomes a brand asset, the feed teaches a subtle lesson: that being okay isn't compelling, that the way to be seen is to be visibly hurting. The predictive model trained on that signal starts to identify suffering as the path to belonging, and the system orients toward states it would otherwise be working to leave.
The other modern pressure here is the pull toward performative public alignment - the implicit demand to visibly endorse whichever social or political stance one's online environment treats as required, with the threat of shaming or status loss for those who don't post or speak up on cue. This pattern operates across the political spectrum; every faction enforces its own version of it. The fear of being seen as not-having-the-right-take recruits the same FEAR-circuit machinery as any other social-exclusion threat, but acts on it continuously - the threat never resolves, because there is always a next stance the algorithm will surface to test alignment against. If you find yourself scrolling and feeling pressured to perform a position to avoid social punishment, that's the predictive model being driven by external cultural signal rather than by what you actually think - and continuing down that path keeps the threat-detection system mobilized against a discrepancy no single action can resolve.
Practically, this means: the chronically anxious person who is on their phone three or four hours per day is feeding the same systems the supplement protocol is trying to recalibrate. Auditing your screen time honestly is comparable to auditing your caffeine intake - many people are habituated and have lost the ability to feel the input as anxiogenic. The protocol that has the best evidence base in self-experimentation: morning phone delay (no phone for the first 30-60 minutes after waking, during which the cortisol awakening response stabilizes), evening phone cutoff (no phone for the last 60-90 minutes before sleep, allowing the parasympathetic transition into sleep onset), specific app deletion or aggressive time limits on the most anxiogenic apps for your particular pattern, and a doomscroll moratorium as a category. Two weeks of this produces effects that often exceed what any single supplement intervention does. The cost is friction, not money.
Present-Moment Practice
Active inference predicts that the predictive model anchored in present sensory evidence generates prediction errors that can be resolved by action, while a model running on memory and imagination generates prediction errors in a space where action is unavailable. Practices that return attention to the present sensory field - body scan meditation, somatic tracking, mindful walking, contemplative prayer, attention to breath - work through this mechanism. They are precision-dial recalibration through evidence-stream redirection.
For the trauma-survivor phenotype specifically, the somatic and parts-based modalities (EMDR, Somatic Experiencing™, IFS™) often have stronger evidence than purely cognitive present-moment practice. The early-life precision-weighting was installed before language was available, and somatic interventions reach that pre-verbal register more directly than cognitive reframing does. If your anxiety has a clear early-life-stress component, prioritize finding a clinician trained in trauma-specific modalities over generic mindfulness practice.
How to supplement
Putting it together
The stack is organized around the mechanisms above: a foundational baseline that brings the system to nutritional sufficiency, HPA-axis support for the stress-reactivity system, and GABAergic and acute-arousal tools for the neurochemical hyperarousal. Weight it to your phenotype and give it a defined window - eight to twelve weeks - with the lifestyle work running in parallel.
These belong under the entire anxiousness protocol regardless of phenotype. Start here before adding the targeted formulas below.
Magnesium is the physiological NMDA receptor antagonist and a positive allosteric modulator of GABA-A receptor sensitivity. Deficiency is the rule rather than the exception in anxious populations, partly because cortisol drives urinary magnesium excretion - the more stressed you are, the more magnesium you lose, and the more anxious the resulting neurochemistry becomes. This is the cycle to interrupt early; the cost is trivial. The glycinate form combines magnesium with glycine, which itself has inhibitory neurotransmitter activity at the spinal cord and brainstem level - the chelate is well-absorbed, well-tolerated, and produces less GI disturbance than oxide or citrate forms.
Reduces the neuroinflammatory cytokine load that sensitizes the descending FEAR circuit and suppresses GABAergic tone. The EPA fraction is particularly relevant for anxiousness and mood - prioritize an EPA-dominant formulation rather than a generic fish oil with an EPA:DHA ratio close to 1:1. The mechanism is less about omega-3 itself than about the inflammation-resolution pathway it feeds: EPA gets converted to specialized pro-resolving mediators (resolvins, protectins) that actively terminate inflammatory signaling rather than simply suppressing it. For inflammatory-driven anxiety phenotypes, SPM Supreme delivers the pre-formed resolution products directly (2-4 softgels daily) and is meaningfully more potent than fish oil for that specific picture.
VDR receptors are present throughout the limbic system and brainstem structures involved in threat detection and stress response. Deficiency is independently associated with anxiety and mood dysregulation, and is prevalent in populations spending most of their daylight hours indoors. K2 (MK-7) is included alongside D3 to direct calcium appropriately during D3 supplementation. Check serum 25-OH-D before and after 90 days; the target range is 50-70 ng/mL, not merely above-deficient (30 ng/mL).
Pyridoxal-5-phosphate is the active cofactor for glutamate decarboxylase - the enzyme that converts glutamate to GABA. B6 deficiency directly impairs GABA synthesis and produces a clinical picture of irritability, anxiety, and poor stress tolerance. P5P is the active form; standard B6 (pyridoxine HCl) requires conversion to P5P, and that conversion can itself be impaired in some genetic and metabolic contexts. If you are going to prioritize one B vitamin, P5P is it for anxiousness. Particularly relevant for GAD1 risk-allele carriers, where supporting GABA synthesis upstream addresses the bottleneck directly.
A broad-spectrum adrenal and HPA support formula built around ashwagandha, eleuthero (Siberian ginseng), and pantothenic acid (vitamin B5). The ashwagandha component - a standardized extract - does the primary work: multiple RCTs demonstrate reductions in serum cortisol, improvements in the cortisol awakening response, and reductions in perceived stress scores over 8-12 weeks of consistent use, primarily through modulation of the hypothalamic CRH signal and downregulation of the cortisol response to subsequent stressors - not cortisol suppression per se, but a normalization of the amplitude and recovery of the HPA response. Pantothenic acid (B5) is a required cofactor for adrenal steroidogenesis (CoA synthesis, needed for cortisol production); the goal is rhythmic cortisol secretion, not elimination, and B5 deficiency impairs the sharp morning spike, contributing to the flat, dysrhythmic curve that characterizes HPA dysregulation.
Eleuthero adds adaptogenic support with a mechanism distinct from ashwagandha - its eleutherosides modulate glucocorticoid receptor binding and support the recovery phase of the stress response rather than blunting the initial response. The combination is well-suited to the exhausted-but-wired phenotype where the goal is normalizing the rhythm. Take in the morning - adaptogens that support the cortisol awakening response are more effective when dosed to align with the morning rise.
Rhodiola occupies a specific niche that distinguishes it from ashwagandha and the other adaptogens: it is best suited to the individual whose anxiousness has a fatigue and depletion component - the person who is simultaneously anxious and exhausted, whose energy is low in the morning, who feels burned through rather than simply wound up. It modulates serotonin and norepinephrine reuptake, supports adrenal resilience, and has consistent evidence for reducing mental fatigue and improving stress tolerance without the sedating profile of the more purely calming adaptogens. The 3% rosavin standardization matters - rosavins are the primary anxiolytic and adaptogenic bioactives, whereas the salidroside fraction tends toward more stimulating effects, so a rosavin-heavy extract is the appropriate choice for anxiousness.
Rhodiola is activating for most people and is best taken in the morning; used in the evening it can delay sleep onset and contribute to the very nocturnal arousal pattern it is meant to help resolve. If your primary presentation is burnout and exhaustion rather than hyperarousal, this is the adaptogen to prioritize. If hyperarousal is primary, ashwagandha (covered in the Adrenal Energy Formula) is the more appropriate starting point.
A liposomal nervous system downregulation formula whose delivery format is as mechanistically relevant as its ingredients: the liposomal phospholipid carrier bypasses first-pass hepatic metabolism and allows absorption directly through the oral mucosa, producing onset within 15-30 minutes rather than the 45-90 minutes typical of capsules. For acute anxious arousal - the elevated evening, the high-stress moment, the pre-sleep window when the mind will not stop - this timing distinction is clinically meaningful. The active components work across multiple GABAergic entry points at once: PharmaGABA (a fermentation-derived GABA that, unlike synthetic GABA, has evidence for crossing the blood-brain barrier and shifting alpha-wave activity) is the primary agent; skullcap (Scutellaria lateriflora) adds GABA-A binding via baicalin; passionflower (Passiflora incarnata) contributes chrysin-mediated partial GABA-A agonism with evidence comparable to low-dose benzodiazepines in small trials; and chamomile flower oil adds apigenin, which binds the benzodiazepine site of the GABA-A receptor - the same site as pharmaceutical anxiolytics, at lower potency and without the dependency profile.
This is the acute-use tool in this stack, taken sublingually 30-60 minutes before the anticipated high-arousal window or as a consistent pre-bed tool - it complements the longer-acting adaptogenic and HPA interventions rather than replacing them.
Built around two primary actives: Sensoril ashwagandha and L-theanine. The Sensoril extract is the withanolide glycoside-standardized form of ashwagandha, with a slightly different phytochemical profile than KSM-66 - it tends toward a more immediately calming effect, which makes it better suited to an acute stress relief formula than a morning adaptogen. L-theanine is an amino acid found in green tea that promotes alpha-wave activity - the relaxed-but-alert state - by modulating glutamate receptor activity and increasing GABA synthesis; its anxiolytic effect is well-documented and dose-dependent between 100-400mg. The combination addresses both the HPA-axis component (Sensoril dampening CRH-driven arousal) and the GABAergic/glutamatergic component (L-theanine modulating the excitatory-inhibitory balance) simultaneously - a well-constructed formula for the person whose anxiousness has both a stress-response and a neurochemical hyperarousal dimension, which is most people. Can be used situationally around high-stress periods or consistently; unlike rhodiola, it can be taken morning or evening depending on the use case.
A targeted formula built around three actives with distinct but complementary mechanisms: zembrin (Sceletium tortuosum extract), L-theanine, and magnesium. Zembrin is the most clinically interesting component - Sceletium tortuosum (kanna) is a South African succulent with a long history of traditional use for mood and stress and an emerging human-trial evidence base. Its mechanism is dual: selective serotonin reuptake inhibition (comparable mechanistically to SSRIs but at much lower potency and without the side-effect profile) and phosphodiesterase-4 (PDE4) inhibition. The PDE4 inhibition is the more pharmacologically interesting - PDE4 breaks down cyclic AMP, a second messenger that modulates the stress response and neuroinflammation, and PDE4 inhibitors are under investigation as antidepressant and anxiolytic agents; zembrin provides this at nutraceutical doses. The magnesium component addresses the NMDA receptor and GABA-A modulation discussed above; the L-theanine overlaps with Enhanced Stress Relief, so note this if combining both and adjust total L-theanine intake (the therapeutic range is 200-400mg; exceeding it does not appear to help).
Appropriate as a consistent daily supplement rather than an acute-use tool given the mechanism of zembrin - the SRI and PDE4 effects build over consistent use. Zembrin's serotonergic activity warrants caution in combination with SSRIs/SNRIs - this is a clinical conversation, not a self-directed combination.
Where to start
How to approach this
The entry point depends on your phenotype. Read the mechanism and genetics sections, identify the primary driver, and start there. In all cases, the foundational baseline (magnesium, omega-3, D3+K2, P5P) goes first - the targeted formulas work much better in a system that has already been brought to nutritional sufficiency.
For HPA-axis-driven anxiety - chronically activated cortisol axis, FKBP5 / NR3C1 / chronic-stress gene-by-environment pattern, sleep disruption common - the entry point depends on which sub-presentation is loudest. The exhausted-but-wired version (wound up despite depletion, evening cortisol elevated, 2-4am wakings) responds to Adrenal Energy Formula's ashwagandha-centered profile, taken in the morning, with Lipocalm for high-arousal windows or as a consistent pre-bed tool. The burned-through version (low energy, low motivation, anxiety that feels more like dread than arousal) responds better to Rhodiola in the morning, where ashwagandha would be too sedating - this sub-presentation often has a dopaminergic-depletion component that overlaps with the depression guide, worth reading in parallel if that resonates. Foundation first in both cases; breathing practice twice daily.
For the hyperarousal, can't-switch-off, cognitively busy pattern where fatigue is less prominent (the slow-clearance COMT ruminator profile): foundation first, then Enhanced Stress Relief covers both the Sensoril HPA dampening and the L-theanine glutamatergic modulation in one formula. Add AnxiaEase for the zembrin PDE4 and amygdala-reactivity mechanism. Cognitive arousal that lives in the head responds particularly well to L-theanine combined with present-moment practice.
For anxiousness with a clear inflammatory or autoimmune component: prioritize SPM Supreme over generic omega-3 in the foundation, address the inflammatory drivers directly (gut work, food sensitivities, autoimmune evaluation), and recognize that the supplement protocol will produce muted effects until the upstream inflammatory load is reduced.
For anxiousness with a clear hormonal component (cyclical, postpartum, perimenopausal): the supplement protocol is not the first line - hormonal evaluation is. Once the hormonal picture is understood and addressed (often through bioidentical progesterone, sometimes estradiol, sometimes thyroid optimization), the supplement stack works as designed. Skipping the hormonal evaluation and going straight to the nutraceutical stack for a hormonally-driven anxiety presentation is a recipe for the partial response that drives so many anxious women through serial supplement protocols without resolution.
For anxiousness with a clear gut component - GI symptoms, food reactivity, the sense that your gut and your nervous system are coupled in a way that makes both worse: the foundational omega-3 and magnesium address the neuroinflammatory side. The gut dysbiosis itself requires more targeted intervention than this guide covers - see A Field Guide to Dysbiosis & SIBO.
For anxiousness with a clear caffeine-sensitivity component (probable ADORA2A risk allele): the most informative single intervention is a two-week caffeine elimination, with the supplement protocol running underneath. Adenosine receptor sensitivity normalizes within 7-14 days of caffeine withdrawal, and the baseline anxiety reduction during that window is often the most dramatic intervention-response relationship in this entire protocol.
For carriers of the BDNF risk variant: exercise becomes a non-negotiable foundation rather than an optional addition. The fear-extinction impairment characteristic of this variant is meaningfully addressed by exercise-induced BDNF upregulation; the supplement stack supports the neurochemistry but doesn't substitute for the BDNF input that movement provides.
Eight to twelve weeks is the evaluation window. HPA axis recalibration, GABA receptor sensitivity normalization, parasympathetic tone recovery, and precision-dial reweighting all operate on that timescale. Lipocalm and Enhanced Stress Relief will produce noticeable acute effects within an hour of dosing. The adaptogenic and structural effects of the rest of the stack are cumulative and require consistent use over months to fully express.
See here for the full supplement stack. We do receive commission on sales from this link/store - if you found this helpful, please support.
The bottom line
The model is what changes
The nervous system that has been running in threat mode for years does not recalibrate in two weeks. It recalibrates through the accumulated evidence that the environment is safe - delivered consistently, through breathing, sleep, movement, present-moment practice, and the biological inputs that tell the system it is no longer under siege. The supplements support that process. They do not replace it. The model is what changes; the supplements lower the cost of the change.
Read the mechanism and the audit, find which are loudest in your picture, and start with the foundation while you do. Weight the targeted formulas to your phenotype, give it eight to twelve weeks, and run the lifestyle work - breathing, sleep, movement, present-moment practice - underneath the whole time. If your picture has a clear hormonal, inflammatory, or gut driver, that evaluation comes first; the stack works as designed once the upstream cause is addressed.
Reference
Appendix A: genetic variant reference
The following is a detailed reference for the genetic variants discussed in this guide. Allele effects, clinical relevance, and mechanistic notes for each SNP.
The Serotonin System
| Gene | Variant | Effect |
|---|---|---|
| SLC6A4 / 5-HTTLPR | rs25531 | Short allele (S/T) = lower serotonin transporter expression, greater amygdala reactivity to threatening stimuli, anxiety-related personality traits, higher rates of generalized anxiety, higher risk of SSRI side effects but stronger response in some carriers. Long allele (L/C) = higher transporter expression, more emotionally stable baseline, less SSRI responsiveness because the transporter is the SSRI target. The most extensively studied genetic variant in anxiety research. |
| 5HT1AR | rs6295 | G allele = higher 5-HT1A autoreceptor expression, stronger serotonin self-inhibition, lower net serotonin output, higher rates of comorbid anxiety-depression, consistently poorer SSRI response. The single most replicated variant in treatment-resistant anxiety-depression presentations. |
| TPH2 | rs4570625, rs11178997 | Risk alleles reduce TPH2 transcriptional activity - slower brain serotonin synthesis at the source, regardless of dietary tryptophan. Relevant to the anxiety phenotype where serotonergic insufficiency is part of the picture. |
The Catecholamine System
| Gene | Variant | Effect |
|---|---|---|
| COMT | rs4680 (Val158Met) | AA (Met/Met) = slow COMT, high baseline dopamine and norepinephrine in PFC, sharp cognition under normal conditions, significantly worse performance under stress, ruminative anxiety phenotype, cognitive overactivation. GG (Val/Val) = fast COMT, lower baseline catecholamines, motivational flatness more than rumination, anhedonic depression risk. The single most actionable catecholamine variant for anxiety phenotyping. |
| MAO-A | promoter VNTR | 3-repeat (low-activity) variant produces slower monoamine breakdown, higher synaptic monoamine concentrations, and increased emotional reactivity. Significant gene-by-environment interaction with early-life stress; behavioral phenotype expression is environment-dependent. |
| DBH | rs1108580 | G allele = higher DBH activity (norepinephrine > dopamine), associated with anxiety phenotype and Parkinson's risk. A allele = lower DBH activity (dopamine > norepinephrine), greater working memory. |
The HPA Axis and Stress Response
| Gene | Variant | Effect |
|---|---|---|
| FKBP5 | rs9296158, rs1360780 | Risk alleles produce glucocorticoid receptor with reduced cortisol negative feedback sensitivity - HPA axis takes longer to terminate after stress activation. Most consistently replicated gene-by-environment interaction in stress-related disorders. Variant + early-life adversity substantially increases anxiety risk. |
| NR3C1 | rs6198, rs10482605 | Risk haplotype reduces glucocorticoid receptor transcriptional activity, increases glucocorticoid resistance, impairs HPA negative feedback. Frequently co-implicated with FKBP5 in chronic anxiety with early-life stress origin. |
| CRHR1 | rs110402 | Affects CRH receptor-1 binding affinity and signal transduction efficiency. Influences amplitude of initial stress response and accompanying anxious arousal. Interacts with FKBP5 and NR3C1 status to determine overall stress reactivity profile. |
The GABAergic and Adenosine Systems
| Gene | Variant | Effect |
|---|---|---|
| GAD1 | rs3791850, rs769390 | Risk variants reduce glutamic acid decarboxylase 1 expression or activity, impairing glutamate-to-GABA conversion and producing relative glutamate excess. Direct genetic mechanism for GABAergic insufficiency. P5P supplementation particularly relevant for risk-allele carriers. |
| GABRA2 | rs279858 | Variants in the GABA-A receptor alpha-2 subunit gene associated with anxiety risk and altered response to benzodiazepines. Less studied than GAD1 but relevant when present alongside other GABA-system variants. |
| ADORA2A | rs5751876 | Adenosine A2A receptor variant. Risk allele associated with markedly increased anxiety, panic, and cardiovascular reactivity in response to caffeine. Carriers may benefit from treating caffeine as a deliberate input rather than a default beverage. |
Neuroplasticity and Fear Extinction
| Gene | Variant | Effect |
|---|---|---|
| BDNF | rs6265 (Val66Met) | Met allele produces reduced activity-dependent BDNF secretion, impairing fear extinction at the molecular level. Met carriers show more persistent anxiety responses, slower extinction of conditioned fear, and weaker response to exposure-based interventions. Exercise as BDNF upregulator carries particular weight for Met carriers. |
| NTRK2 | rs1212171 | Encodes the BDNF receptor TrkB. Variants modulate the cellular response to BDNF signaling. Less actionable than BDNF Val66Met but informative when BDNF response appears blunted despite adequate baseline. |
This guide is educational and is not a substitute for individualized care from a licensed healthcare provider. Nothing here is a diagnosis or a prescription. Several of the supplements below interact with prescription medications - zembrin and L-theanine alongside SSRIs/SNRIs, magnesium with certain medications, ashwagandha in thyroid and autoimmune disease - so talk to your physician or a functional-medicine clinician before starting, especially if you take prescription medication, are pregnant or breastfeeding, or manage a chronic condition. Do not discontinue a benzodiazepine or an SSRI to begin this protocol. Merlin may earn a commission on products purchased through the Fullscript plan linked here.