Health Updated January 2026

How to manage anxiety

Anxiety advice is everywhere—most of it vague or contradictory. We reviewed the research and tracked what clinicians and neuroscientists actually agree on, then built a practical toolkit from the evidence.

Key finding

The evidence points to a few things that actually work

After reviewing dozens of meta-analyses and tracking six leading researchers, the interventions with strongest evidence are: cognitive behavioral therapy (CBT), regular exercise, sleep optimization, and specific breathing techniques. Most "biohacks" have weaker support, though some supplements show promise.

Note: This guide covers general anxiety management. If you're experiencing severe anxiety, panic attacks, or thoughts of self-harm, please consult a mental health professional. This information supplements but doesn't replace professional care.

Researchers we track

Andrew Huberman

Stanford neuroscientist

David Burns

CBT pioneer, Stanford

Judson Brewer

Anxiety researcher, Brown

Peter Attia

Longevity physician

Matthew Walker

Sleep researcher, UC Berkeley

Anna Lembke

Stanford Addiction Medicine

We chose these six because they hold academic positions, cite primary research, and have accessible public content (podcasts, books) that translates research into actionable advice.

What the evidence strongly supports

Interventions with multiple meta-analyses and consistent research support

Cognitive behavioral therapy (CBT)

Very strong evidence

The consensus: CBT is the most studied psychological treatment for anxiety, with effect sizes of 0.24-0.56 across meta-analyses. Benefits are maintained at 6-12 months post-treatment—often better than medication for long-term outcomes.

Practical implementation

  • • Find a CBT-trained therapist (Psychology Today directory filters by approach)
  • • If cost is a barrier: David Burns' Feeling Good is clinically validated for self-help
  • • Apps like Woebot or MindShift offer guided CBT exercises
  • • Typical course: 8-16 weekly sessions

Key sources:

• Hofmann & Smits (2018): 41 RCTs, n=2,843 — effect size g=0.56 vs placebo

• Long-term outcomes meta-analysis (2019): Benefits maintained at 6-12 months

• Peter Attia's Episode #362 with Josh Spitalnick on CBT for anxiety

Regular exercise

Strong evidence

The consensus: Exercise reduces anxiety symptoms with effect sizes of -0.39 to -0.60—comparable to medication for some populations. Physically active people have 34% lower odds of developing anxiety disorders.

Practical implementation

  • • Any movement counts—walking, swimming, weights, yoga
  • • Target: 150 min/week moderate or 75 min vigorous
  • • Resistance training may be particularly effective for anxiety
  • • Consistency matters more than intensity—start small

Key sources:

• BJSM Overview (2023): Physical activity effect size -0.42 for anxiety

• Prospective cohort meta-analysis (2019): 34% lower anxiety risk in active individuals

• Older adults meta-analysis (2024): Effect size -0.60

Sleep optimization

Strong evidence

The consensus: The sleep-anxiety relationship is bidirectional and powerful. People with insomnia are 17x more likely to have clinical anxiety and 20x more likely to have panic disorder. Improving sleep often improves anxiety—and vice versa.

Practical implementation

  • • Consistent wake time (±30 min), even on weekends
  • • 7-9 hours opportunity for sleep
  • • No screens 1 hour before bed (or use night mode)
  • • Cool, dark room (65-68°F / 18-20°C)
  • • If insomnia persists: CBT-I is more effective than sleeping pills long-term

Key sources:

• Stanford Medicine Review (2025): 17x anxiety risk with insomnia

• Matthew Walker's Drive interviews on sleep and mental health

• See our energy guide for detailed sleep protocols

Breathing techniques (especially cyclic sighing)

Strong evidence

The consensus: A 2023 Stanford RCT found that 5 minutes of cyclic sighing daily was more effective than mindfulness meditation for improving mood and reducing anxiety. This is the most time-efficient intervention we've found.

How to do cyclic sighing (5 min/day)

  • 1. Inhale through your nose
  • 2. Take a second, deeper inhale to fully expand your lungs
  • 3. Exhale slowly and completely through your mouth
  • 4. Repeat for 5 minutes

For acute anxiety: physiological sigh

  • • Same technique, but just 1-3 breaths
  • • Works in real-time to reduce stress response
  • • Use before presentations, difficult conversations, etc.

Key sources:

• Stanford cyclic sighing RCT (2023): n=111, cyclic sighing > meditation for mood

• Huberman Lab Tools for Managing Stress & Anxiety

• Stanford Medicine coverage

What has moderate evidence

Promising interventions with good but not definitive research support

Mindfulness meditation

Moderate evidence

The evidence: A 2014 JAMA review of 47 trials found moderate evidence for anxiety reduction (effect size 0.38). Similar efficacy to CBT, but notably—not superior to active treatments like exercise or medication.

Practical implementation

  • • Start with 10 min/day using an app (Headspace, Calm, Waking Up)
  • • Focus on breath awareness or body scan
  • • Consistency matters more than duration
  • • Note: cyclic sighing is faster and has stronger evidence

Key sources:

• Goyal et al. (2014): JAMA review, 47 trials, effect size 0.38

Caffeine reduction

Moderate evidence

The evidence: 51% of panic disorder patients experienced a panic attack after caffeine (vs 0% with placebo). Individual sensitivity varies enormously—some people tolerate caffeine fine, others are highly reactive.

Practical implementation

  • • If you have panic attacks: try eliminating caffeine for 2 weeks
  • • General anxiety: try reducing to under 200mg/day (2 cups coffee)
  • • No caffeine after early afternoon regardless (protects sleep)
  • • Taper gradually to avoid withdrawal headaches

Key sources:

• Caffeine and panic meta-analysis (2021): 51% panic rate in PD patients

Supplements

Moderate evidence

The evidence: Some supplements show promise in clinical trials, but the research is less robust than for CBT or exercise. Use as adjuncts, not replacements for evidence-based treatments.

L-Theanine

100-200mg · Fast-acting

Increases alpha waves, reduces cortisol. Good for acute situational anxiety. Found in tea.

Ashwagandha

300-600mg · Takes weeks

Reduces cortisol over time. WFSBP provisionally recommends for GAD. Monitor thyroid if long-term use.

Magnesium glycinate

200-400mg · Evening

Many people are deficient. May help with sleep and muscle tension. Glycinate form is well-absorbed.

Omega-3 (EPA/DHA)

1-2g · With food

Anti-inflammatory effects may help anxiety. Most people don't get enough from diet.

Key sources:

• L-Theanine RCT: Reduces cortisol, increases alpha waves

• NIH Ashwagandha fact sheet: 7 studies, reduces stress and cortisol

What has emerging evidence

Interesting but requires more research

Cold exposure

Emerging

Cold water immersion triggers dopamine and norepinephrine release, creating acute increases in alertness and potentially reducing anxiety. A 2025 PLOS ONE review found significant stress reduction, but studies are small and can't be blinded.

Our take: Worth trying if curious. The acute mood boost is real. Long-term anxiety benefits are less certain. Start with cold showers, not ice baths.

Social connection

Strong for general MH

The evidence for social connection and overall mental health is very strong (lonely people are 2x more likely to develop depression). But the causal relationship with anxiety specifically is complex—social anxiety can cause isolation, which worsens anxiety.

Our take: Prioritize in-person connection over social media. One study found limiting social media to 10 min/day significantly reduced anxiety. Quality over quantity.

Where experts disagree

Areas of legitimate debate in the research community

Medication vs. therapy as first-line treatment

Current guidelines: Both SSRIs and CBT are considered equally valid first-line options. Patient preference should guide the choice.

The debate: Medication works faster initially, but CBT may have more durable long-term effects. Some argue therapy should be tried first since it has no side effects.

Our take: Neither is universally "better." If you need faster relief or can't access good CBT, medication is reasonable. If you prefer non-pharmacological approaches and can commit to weekly sessions, start with CBT. Combination therapy may be most effective for severe anxiety.

Exposure therapy approach: habituation vs. inhibitory learning

Traditional view: Fear must decrease during exposure for learning to occur.

Newer view (Michelle Craske, UCLA): The original fear isn't erased—a new, competing memory is formed. Expectancy violation matters more than fear reduction.

Why it matters: If you're doing exposure therapy and don't feel less anxious during sessions, that doesn't mean it's not working. The inhibitory learning model suggests maximizing surprise and varying contexts may improve outcomes.

The anxiety toolkit: where to start

If you're overwhelmed by options, here's our suggested priority order based on evidence strength and effort required.

1

Cyclic sighing (5 min/day)

Highest ROI intervention. Start today.

2

Fix your sleep

17x anxiety risk with insomnia. Non-negotiable foundation.

3

Move your body

Any exercise helps. Walking counts. 150 min/week target.

4

Cut caffeine (if prone to panic)

51% panic rate in susceptible individuals. Try 2 weeks without.

5

Consider CBT

If anxiety persists, this is the gold standard. Find a trained therapist or start with Feeling Good.

Primary sources

The researchers, clinicians, and key works we track for this guide

Huberman Lab: Stress & Anxiety

Episode with protocols

David Burns: Feeling Good

CBT self-help (10M+ copies)

Judson Brewer: Unwinding Anxiety

Mindfulness-based approach

The Drive: Episode #362

CBT for anxiety deep-dive

Key research cited

Change log

  • January 2026: Initial publication