By Dr. Vishal Saini, M.D., FAASM — Mid-West Center for Sleep Disorders
In an earlier article I laid out the evidence-based, non-drug ways to protect and modestly increase N3 — slow-wave sleep, the deepest, most physically restorative stage. The reasonable next question, and one I hear often, is: "I've done all that. Isn't there a medication that just increases my deep sleep?"
This piece is my honest answer, written the way I'd reason through it in clinic. It's a review of what pharmacology can and can't do for N3, organized by drug class. And I'll say the punchline up front, because it matters: there is no FDA-approved medication whose job is to increase deep sleep in an otherwise healthy adult, most common sleeping pills do not increase N3 — and some actively suppress it — and the drugs that genuinely do boost slow-wave sleep are, almost without exception, either reserved for specific diseases, held back by side effects, or a cautionary tale about why "more N3" is not the same as "better sleep."
Let me show you the reasoning.
First, What "Conservative Measures Failed" Should Actually Mean
Before any conversation about medication, I want to be sure we're solving a real problem, because "my tracker says my deep sleep is low" is not, by itself, a diagnosis or a valid drug indication.
Genuinely exhausting the conservative options means: a real course of cognitive behavioral therapy for insomnia (CBT-I) — the strongly recommended first-line treatment for chronic insomnia (Edinger et al., J Clin Sleep Med, 2021, PMID 33164742) — not just a printout of sleep-hygiene tips. It means a consistent schedule and morning light, regular exercise, and sensible pre-bed temperature and substance habits. And critically, it means we've ruled out and treated the things that fragment deep sleep in the first place: untreated obstructive sleep apnea, restless legs and periodic limb movements, depression, chronic pain, and circadian misalignment. A patient whose N3 is being shredded by unrecognized apnea does not need a deep-sleep drug; they need their apnea treated.
Only once that groundwork is truly done does the pharmacology question become legitimate. And even then, the goal is rarely "raise the N3 number" — it's "help this specific person sleep and feel better." Keep that distinction in mind, because it's the thread running through everything below.
"Patients ask me for a deep-sleep pill the way they'd ask for a supplement. But slow-wave sleep isn't a deficiency you top up. If it's genuinely low, the useful question is why — and the answer is usually a treatable disorder, not a missing drug." — Dr. Vishal Saini
The Uncomfortable Headline: The Approved Insomnia Drugs Don't Target Depth
Here's what surprises people most. The American Academy of Sleep Medicine's clinical practice guideline for the pharmacologic treatment of chronic insomnia evaluated the common prescription options and made only weak recommendations, and — this is the key part — every one of them is aimed at sleep onset or sleep maintenance, not sleep depth (Sateia et al., J Clin Sleep Med, 2017, PMID 27998379). The suggested-for list includes agents like eszopiclone, zolpidem, temazepam, triazolam, zaleplon, ramelteon, doxepin, and suvorexant — for falling asleep or staying asleep. None is recommended because it increases N3.
Just as telling, the same guideline recommends against several agents that people specifically associate with "deeper" sleep — including trazodone, tiagabine, diphenhydramine, melatonin, and valerian — because the evidence didn't support a favorable benefit-to-risk balance for insomnia (Sateia et al., 2017, PMID 27998379). We'll come back to why tiagabine's appearance on the "against" list is so instructive, given that it demonstrably increases slow-wave sleep.
So the honest starting point is that our approved toolkit was never built to increase deep sleep. It was built to help people sleep.
The Paradox: Many Sleep Drugs Reduce Deep Sleep
This is the part that should give anyone chasing N3 with medication real pause.
Benzodiazepines (temazepam, triazolam, and the like) are classic examples: by their action on the GABA-A receptor they tend to increase light stage N2 and decrease slow-wave sleep, while also suppressing REM. They can make you sleep longer and feel knocked out, but they generally move your architecture in exactly the wrong direction if your goal is more deep sleep. This is well established in sleep pharmacology and is one reason I avoid them for this purpose entirely.
The "Z-drugs" (zolpidem, zaleplon, eszopiclone) act at the same receptor family and are, at best, relatively neutral toward slow-wave sleep — not the reliable enhancers they're sometimes assumed to be. And older sedating antihistamines like diphenhydramine (the "PM" in over-the-counter sleep aids) bring tolerance within days and anticholinergic risks that make them a poor long-term choice, which is why the guideline advises against them (Sateia et al., 2017, PMID 27998379).
The lesson: "take a sleeping pill to get deeper sleep" is often self-defeating. Sedation is not the same as slow-wave sleep — a point I keep returning to because patients so reasonably conflate the two.
The Drugs That Genuinely Increase Slow-Wave Sleep — and Their Catches
Now to the agents that do raise N3. Each comes with a serious asterisk.
Oxybate (sodium oxybate / low-sodium oxybate) — the most potent, and the most restricted
If any drug can be said to "increase deep sleep," it's oxybate. Taken at night, it profoundly consolidates sleep and increases slow-wave sleep and delta (slow-wave) EEG activity — an effect demonstrated in controlled polysomnography studies (Walsh et al., Sleep, 2010, PMID 20857869) and a large part of why it improves the unrefreshing sleep and daytime sleepiness of narcolepsy. It's approved for narcolepsy (and, in its newer lower-sodium form, for idiopathic hypersomnia), and it's genuinely effective there.
But oxybate is not a general-purpose deep-sleep drug, and it shouldn't be. It's a potent central nervous system depressant (chemically a close relative of GHB), tightly controlled through a restricted distribution program because of abuse and safety concerns, dangerous in combination with alcohol or other sedatives, and dosed in the middle of the night. Its risk-benefit profile makes sense for a disabling hypersomnia disorder; it makes no sense to chase a better N3 percentage in someone who is otherwise well. It's the clearest proof that we can pharmacologically increase deep sleep — and that doing so is only justified for the right disease.
Gabapentinoids (gabapentin, pregabalin) — real SWS effects, best when a comorbidity fits
Gabapentin and pregabalin increase slow-wave sleep in polysomnography studies (pregabalin: Bazil et al., Epilepsy & Behavior, 2012, PMID 22424859; gabapentin: Susantitapong et al., Clinical and Translational Science, 2024, PMID 38803031), which is one reason they can genuinely improve sleep quality. In practice they're most useful when the patient's insomnia travels with something these drugs also treat — neuropathic or other chronic pain, restless legs syndrome, or prominent anxiety. In that setting, deepening sleep is a welcome bonus of treating the comorbidity, which is exactly the right way to use them.
The catches are real, though: daytime sedation and dizziness, fluid retention and weight gain, misuse potential (increasingly recognized with gabapentinoids), and withdrawal if stopped abruptly. These are off-label for primary insomnia, and I reserve them for patients whose overall picture makes them a sensible fit — not as a deep-sleep booster in isolation.
Tiagabine and gaboxadol — the cautionary tale
Here's where the story turns instructive. Tiagabine, an anti-seizure drug that raises synaptic GABA, increases slow-wave sleep in a dose-dependent fashion in sleep-lab studies (Taranto-Montemurro et al., Sleep, 2017, PMID 28364504). On paper, that's exactly what a "deep sleep drug" should do. Yet it did not translate into meaningful benefit for insomnia patients, carried its own risks, and is specifically among the agents the AASM guideline recommends against (Sateia et al., 2017, PMID 27998379). More pointedly, when researchers used tiagabine to boost slow-wave sleep in volunteers, the extra deep sleep did not improve overnight memory consolidation (Feld et al., Sleep, 2013, PMID 23997364) — the deep sleep went up, and the thing it was supposed to help did not.
Gaboxadol tells a similar story from the other direction: a compound developed specifically as a slow-wave-sleep enhancer that reliably increased N3 in trials (Walsh et al., Sleep, 2008, PMID 18517036) — and whose development as an insomnia drug was nonetheless abandoned after the overall results and tolerability didn't justify it.
Together these two make the single most important point in this entire review: increasing the N3 number is not the same as helping the patient. Two drugs did precisely what the "more deep sleep" logic asks for, and neither earned a place in insomnia care. Any pitch — from a supplement, a device, or a prescriber — that treats a higher deep-sleep figure as self-evidently good is ignoring this lesson.
Trazodone and other sedating antidepressants
Trazodone at low doses is one of the most commonly prescribed off-label sleep aids in the country, and it does have measurable effects on sleep architecture, including on slow-wave sleep (a systematic review and meta-analysis; Zheng et al., Scientific Reports, 2022, PMID 36002579). But the evidence base for it in primary insomnia is thin enough that the AASM guideline recommends against it as well (Sateia et al., 2017, PMID 27998379), and it carries its own issues (next-day grogginess, dizziness, rare priapism). It has a legitimate role — particularly when insomnia coexists with depression, where an antidepressant is independently indicated — but it is not a validated deep-sleep treatment.
Mirtazapine likewise increases slow-wave sleep and is reasonable when depression and insomnia (and low appetite/weight) coexist, at the cost of sedation and weight gain. Low-dose doxepin is FDA-approved for sleep-maintenance insomnia and is notable for the opposite reason — it improves sleep continuity through selective histamine blockade with relatively little effect on sleep architecture, i.e., it doesn't meaningfully change N3. Useful drug; just not a deep-sleep drug.
Dual orexin receptor antagonists (daridorexant, lemborexant, suvorexant) — the modern, architecture-sparing approach
The newest class is worth understanding because it reflects where the field has sensibly moved. Rather than force sedation through the GABA system, dual orexin receptor antagonists (DORAs) dial down wake-promoting orexin signaling. Their appeal isn't that they dramatically inflate N3 — they don't — but that they promote sleep while largely preserving normal sleep architecture (daridorexant: Yagi et al., Sleep and Biological Rhythms, 2026, PMID 41969977), in contrast to the slow-wave suppression seen with benzodiazepines. Suvorexant is among the agents the AASM guideline suggests for sleep maintenance (Sateia et al., 2017, PMID 27998379), and the newer members of the class were designed with next-day function in mind.
For a patient who needs pharmacologic help and in whom I want to avoid degrading deep sleep, a DORA is often a more rational choice than an older hypnotic — not because it boosts N3, but because it doesn't wreck it.
5-HT2A antagonists (investigational) — another "more isn't better" story
For completeness: blocking the serotonin 5-HT2A receptor robustly increases slow-wave sleep, and several such compounds (the older agent ritanserin and later inverse agonists like pimavanserin) demonstrated exactly that on polysomnography (Ancoli-Israel et al., Sleep Medicine, 2011, PMID 21256805). None became a marketed insomnia therapy. Once again, a clean pharmacologic increase in N3 did not deliver the clinical outcomes that would justify bringing a drug to patients.
Agents to specifically avoid for this purpose
Sedating antipsychotics such as quetiapine and olanzapine do increase slow-wave sleep, and they are unfortunately prescribed off-label for sleep. I want to be unambiguous: their metabolic, cardiac, and movement-disorder risks make them inappropriate for insomnia or for chasing deep sleep, outside of a genuine psychiatric indication. A better N3 number is not worth those risks.
Putting It Together: How I Actually Approach This in Clinic
If a patient has truly optimized the conservative measures and still isn't sleeping or feeling well, here's the logic I follow — and notice that "raise N3" is never the target itself:
- Re-confirm the groundwork. Was CBT-I actually completed? Is there untreated apnea, restless legs, depression, pain, or a circadian problem quietly driving this? Treat those first; deep sleep usually follows.
- Match any medication to the comorbidity, not to a deep-sleep goal. Restless legs or neuropathic pain alongside the insomnia? A gabapentinoid may help sleep and the comorbidity together. Depression? An antidepressant with sedating properties may be justified on its own merits. A diagnosed central hypersomnia like narcolepsy? That's the specific, appropriate place for oxybate.
- If a pure hypnotic is warranted, prefer agents that don't degrade architecture. For sleep-maintenance trouble, a DORA is often my preference over a benzodiazepine precisely because it spares slow-wave and REM sleep.
- Avoid the traps. No benzodiazepines or antihistamines as a "deep sleep" strategy; no antipsychotics for insomnia; and no treating a wearable's deep-sleep percentage as a drug target.
The through-line: we treat people and disorders, not numbers on a hypnogram.
What the Research Still Hasn't Settled
In fairness, this is an area with real open questions. We do not have good evidence that pharmacologically inflating slow-wave sleep, on its own, improves the outcomes patients actually care about — daytime alertness, memory, mood, cardiometabolic and brain health — in people without a defined sleep disorder. The tantalizing links between slow-wave sleep and overnight brain "clearance" have driven interest in SWS-enhancing drugs for conditions like Alzheimer's disease, but that remains investigational, not established therapy. And the failures of gaboxadol, tiagabine, and the 5-HT2A compounds as insomnia drugs are a standing reminder that the field has repeatedly increased N3 without improving lives. Until a drug shows it improves outcomes, not just the EEG, "boost your deep sleep" should be read as a hypothesis, not a prescription.
The Bottom Line
There is no approved medication to increase deep sleep in an otherwise healthy adult, and reaching for one is usually the wrong move. Most common sleeping pills don't raise N3, and benzodiazepines actively lower it. The drugs that genuinely increase slow-wave sleep — oxybate above all, then gabapentinoids, and the cautionary cases of tiagabine, gaboxadol, and the 5-HT2A antagonists — are either reserved for specific diseases, limited by side effects, or living proof that more N3 is not automatically better sleep. When conservative measures fail, the answer is rarely a deep-sleep drug; it's a careful search for the treatable reason your sleep isn't deep, and treatment matched to that. If a medication belongs in your plan, it should earn its place by treating a real disorder — and, ideally, by protecting your sleep architecture rather than distorting it.
Dr. Vishal Saini, M.D., FAASM is the Research & Medical Director at Mid-West Center for Sleep Disorders and Principal Investigator on multiple clinical trials in sleep medicine across Michigan. He evaluates and treats insomnia, sleep apnea, narcolepsy, and complex hypersomnia disorders in Lansing, Traverse City, and Eaton Rapids.
Tried everything and still not sleeping deeply? The answer is usually a diagnosis, not a stronger pill. Let's find it.
Book a Consultation → | (517) 887-6733
References: Sateia MJ, Buysse DJ, Krystal AD, Neubauer DN, Heald JL. Clinical practice guideline for the pharmacologic treatment of chronic insomnia in adults: an AASM clinical practice guideline. J Clin Sleep Med 2017 (PMID 27998379); Edinger JD, Arnedt JT, Bertisch SM, et al. Behavioral and psychological treatments for chronic insomnia disorder in adults: an AASM clinical practice guideline. J Clin Sleep Med 2021 (PMID 33164742); Walsh JK, Hall-Porter JM, Griffin KS, et al. Enhancing slow wave sleep with sodium oxybate reduces the behavioral and physiological impact of sleep loss. Sleep 2010 (PMID 20857869); Bazil CW, Dave J, Cole J, et al. Pregabalin increases slow-wave sleep and may improve attention in patients with partial epilepsy and insomnia. Epilepsy Behav 2012 (PMID 22424859); Susantitapong K, Dilokpattanamongkol P, Sutherasan Y, et al. Effects of gabapentin on slow-wave sleep period in critically ill adult patients: a randomized controlled trial. Clin Transl Sci 2024 (PMID 38803031); Taranto-Montemurro L, Sands SA, Edwards BA, et al. Effects of tiagabine on slow wave sleep and arousal threshold in patients with obstructive sleep apnea. Sleep 2017 (PMID 28364504); Feld GB, Wilhelm I, Ma Y, et al. Slow wave sleep induced by GABA agonist tiagabine fails to benefit memory consolidation. Sleep 2013 (PMID 23997364); Walsh JK, Snyder E, Hall J, et al. Slow wave sleep enhancement with gaboxadol reduces daytime sleepiness during sleep restriction. Sleep 2008 (PMID 18517036); Zheng Y, Lv T, Wu J, Lyu Y. Trazodone changed the polysomnographic sleep architecture in insomnia disorder: a systematic review and meta-analysis. Sci Rep 2022 (PMID 36002579); Yagi T, Ozone M, Ioji T, et al. Effects of daridorexant on sleep architecture in Japanese patients with insomnia disorder: analysis of a phase II randomized controlled trial. Sleep Biol Rhythms 2026 (PMID 41969977); Ancoli-Israel S, Vanover KE, Weiner DM, et al. Pimavanserin tartrate, a 5-HT2A receptor inverse agonist, increases slow wave sleep as measured by polysomnography in healthy adult volunteers. Sleep Med 2011 (PMID 21256805).