For decades, the hardest part of schizophrenia hasn’t been the hallucinations or delusions. Those symptoms — the dramatic ones, the ones depicted in films and whispered about in waiting rooms — respond, at least partially, to antipsychotic drugs that have been on the market since the 1950s. The real clinical nightmare is something quieter and far more disabling: cognitive decline. Memory loss. Fractured attention. An inability to plan, organize, or think abstractly. These deficits strip patients of their independence more completely than any voice they might hear, and until now, no approved drug has touched them.
That may be about to change.
A team of researchers at Northwestern University has identified a new biomarker and a promising drug candidate that directly targets cognitive symptoms in schizophrenia — a first in a field littered with failed attempts. Their findings, published in March 2026 and reported by Northwestern News, center on an enzyme called HDAC5 that appears to be dramatically overexpressed in the brains of people with schizophrenia. When the researchers blocked this enzyme in animal models, cognitive function improved. Significantly.
The study, led by Dr. Peter Bhatt, an assistant professor of psychiatry and behavioral sciences at Northwestern University Feinberg School of Medicine, represents years of painstaking work that began with a simple but vexing question: why do the brains of schizophrenia patients seem to lose their cognitive architecture even when psychotic symptoms are controlled?
“Cognitive symptoms are the number one predictor of long-term functional outcomes in schizophrenia,” Bhatt told Northwestern News. “Yet we have no FDA-approved treatments for them. That’s a massive unmet need.”
Massive is right. Schizophrenia affects roughly 24 million people worldwide, according to the World Health Organization. The cognitive deficits associated with the disorder typically emerge before the first psychotic episode and persist — often worsen — over a patient’s lifetime. They are the primary reason most people with schizophrenia can’t hold jobs, live independently, or maintain relationships. Antipsychotics blunt hallucinations. They do essentially nothing for cognition. In some cases, they make it worse.
The Northwestern team’s approach was different from the start. Rather than screening thousands of compounds hoping one might work, they went looking for the biological mechanism driving cognitive failure. They analyzed postmortem brain tissue from patients with schizophrenia and compared it with tissue from healthy controls. What they found was striking: levels of HDAC5, a histone deacetylase enzyme involved in gene regulation, were abnormally elevated in the prefrontal cortex — the brain region most critical for higher-order thinking, working memory, and executive function.
Histone deacetylases are molecular editors. They modify how tightly DNA is wound around histone proteins, which in turn controls whether certain genes get switched on or off. When HDAC5 is overactive, it effectively silences genes that neurons need to form and maintain synaptic connections. The result: weakened neural circuits in precisely the brain regions responsible for cognition.
So the team designed a selective HDAC5 inhibitor. Not a broad-spectrum HDAC blocker — those have been tried in cancer treatment and come with brutal side effects — but a compound engineered to hit HDAC5 specifically while leaving other family members alone. Selectivity matters enormously here. The HDAC family has 18 members, and they perform wildly different functions throughout the body. Shut down the wrong ones and you get toxicity. Shut down the right one and you might restore cognitive function without the collateral damage that has doomed earlier attempts.
In mouse models of schizophrenia-like cognitive impairment, the selective HDAC5 inhibitor restored performance on tasks measuring working memory and cognitive flexibility. The animals treated with the compound performed comparably to healthy controls. Animals that didn’t receive treatment continued to struggle. The drug didn’t appear to affect psychotic-like behaviors one way or the other, which is actually encouraging — it suggests a clean mechanism of action targeting cognition specifically, not a blunt pharmacological instrument.
“This gives us both a biomarker to identify patients who might benefit and a therapeutic strategy to help them,” Bhatt said in the Northwestern News report. The dual nature of the discovery — diagnostic and therapeutic — is what makes it particularly compelling to researchers in the field.
The biomarker angle shouldn’t be underestimated. One of the persistent problems in psychiatric drug development is heterogeneity. Schizophrenia isn’t one disease. It’s likely many diseases with overlapping symptoms, driven by different underlying biology in different patients. Clinical trials for cognitive enhancers in schizophrenia have repeatedly failed, and one major reason is that researchers have been testing drugs on mixed populations — some of whom may have the relevant biological deficit and some who don’t. If HDAC5 overexpression can be measured in living patients (through blood-based proxies, cerebrospinal fluid, or advanced neuroimaging), it becomes possible to select the right patients for the right trial. That alone could transform the odds of success.
The pharmaceutical industry has been burned badly in this space. Companies including Pfizer, Roche, and Eli Lilly have all invested heavily in cognitive enhancement programs for schizophrenia over the past two decades, and all have walked away with little to show for it. The MATRICS initiative, launched by the National Institute of Mental Health in the early 2000s, was designed to create a standardized framework for developing and testing cognitive treatments in schizophrenia. It produced consensus on measurement tools and clinical trial design. What it didn’t produce was a single approved drug.
The failures weren’t for lack of trying. Glutamate modulators, nicotinic receptor agonists, PDE10A inhibitors — all showed promise in preclinical models and early-phase trials, then collapsed in Phase II or Phase III. The reasons varied. Some compounds didn’t cross the blood-brain barrier effectively in humans. Some worked in rodents but not in primates. Some ran into the heterogeneity problem described above. And some simply targeted the wrong mechanism.
What distinguishes the Northwestern work is the strength of the human tissue data. The team didn’t start with a drug and go looking for a disease. They started with diseased human brains, identified a molecular abnormality, validated it across multiple samples, and then built a therapeutic strategy around it. That sequence — from patient to mechanism to drug — is the direction psychiatric research has been trying to move for years but has rarely achieved in practice.
There are caveats, of course. Mouse models of schizophrenia are imperfect approximations of a uniquely human disorder. No rodent hears accusatory voices or develops paranoid delusions about government surveillance. The cognitive tasks used in mice — set-shifting, novel object recognition, T-maze alternation — capture certain aspects of executive function but can’t replicate the full complexity of human cognition. And the leap from a promising preclinical compound to an FDA-approved medication is long, expensive, and statistically unlikely to succeed. Roughly 90% of drugs that enter clinical trials never reach the market.
But the field is paying attention. The study arrives at a moment when interest in epigenetic approaches to psychiatric disease is surging. Epigenetics — the study of how gene expression is modified without changes to the underlying DNA sequence — has become one of the most active areas of neuroscience research. HDAC inhibitors are already approved for certain cancers (vorinostat for cutaneous T-cell lymphoma, for instance), which means the basic pharmacology is understood and manufacturing pathways exist. Developing a selective HDAC5 inhibitor for a psychiatric indication would be novel, but it wouldn’t require building an entirely new drug class from scratch.
The timing is also notable because of recent shifts in how the FDA evaluates psychiatric drugs. In 2023, the agency approved the first new mechanism antipsychotic in decades — KarXT (xanomeline-trospium), developed by Karuna Therapeutics and later acquired by Bristol-Myers Squibb. That approval, for the treatment of schizophrenia, signaled a willingness to embrace new pharmacological approaches after years of therapeutic stagnation. A drug that specifically targets cognitive symptoms — rather than psychosis — would represent an even more dramatic departure from the status quo, and regulatory conversations about how to design and evaluate such trials are already underway.
The economic implications are substantial. Schizophrenia costs the U.S. healthcare system an estimated $281 billion annually, according to a 2024 analysis published in the journal Schizophrenia Research. The majority of those costs aren’t driven by hospitalizations for acute psychosis. They’re driven by chronic disability — the inability of patients to work, the need for supported housing, the long-term burden on caregivers and social services. A drug that meaningfully improved cognitive function could shift patients from dependency to at least partial self-sufficiency. Even a modest effect size, applied across millions of patients, would translate into billions in reduced societal costs.
Bhatt’s team is now working to refine the HDAC5 inhibitor for human use and develop a blood-based biomarker assay that could be deployed in clinical settings. The goal is to move into Phase I safety trials within the next few years, though specific timelines haven’t been disclosed. Northwestern has filed patent applications related to the compound and the biomarker technology.
Other research groups are watching closely. Dr. Daniel Bhatt (no relation) at Mount Sinai’s Icahn School of Medicine, who studies epigenetic mechanisms in psychiatric disorders, has noted in previous interviews that HDAC5 sits at a particularly interesting intersection of stress biology and cognitive function. The enzyme is known to be activated by chronic stress, which is itself a major risk factor for schizophrenia onset and cognitive decline. This raises an intriguing possibility: HDAC5 overexpression in schizophrenia may not just be a downstream consequence of the disease but could be part of the causal chain linking environmental stress to cognitive deterioration.
If that’s true, an HDAC5 inhibitor might have applications beyond schizophrenia. Cognitive deficits are prominent features of bipolar disorder, major depression, and PTSD. They’re also central to the clinical picture in early Alzheimer’s disease. A selective compound that restores synaptic plasticity in the prefrontal cortex could, in theory, benefit patients across multiple diagnostic categories — though that kind of cross-indication potential is speculative at this stage.
The broader scientific context matters too. Schizophrenia genetics has exploded in the past decade. The Psychiatric Genomics Consortium’s landmark 2022 study identified 287 genetic loci associated with schizophrenia risk — a dramatic expansion from earlier genome-wide association studies. Several of those loci implicate chromatin remodeling and histone modification pathways, which is exactly the biology HDAC5 operates in. The Northwestern finding doesn’t exist in isolation. It sits within a growing body of evidence pointing to epigenetic dysregulation as a core feature of the disorder.
None of this guarantees success. The history of neuroscience is full of beautiful mechanistic stories that fell apart when drugs entered human trials. Biology is messy. Brains are complicated. And schizophrenia, with its tangled web of genetic, developmental, and environmental contributors, remains one of the most formidable challenges in all of medicine.
But here’s what’s different this time: the starting point is human pathology, not a theoretical model. The target was found in real patient brains, not inferred from animal experiments. The biomarker offers a way to stratify patients, which could prevent the kind of underpowered, heterogeneous trials that have killed so many previous candidates. And the drug is designed with selectivity that earlier HDAC approaches lacked.
For the 24 million people living with schizophrenia worldwide — and for the families, clinicians, and payers who bear the consequences of untreatable cognitive decline — even cautious optimism feels like a lot. It’s been a long time since this field had a genuinely new idea that held together from bench to bedside planning. Whether the HDAC5 story ultimately delivers a drug or simply opens a new line of investigation, it has already accomplished something rare: it has given researchers a credible molecular target for the symptom domain that matters most and has resisted treatment the longest.
The next few years will determine whether that target holds up under the unforgiving scrutiny of human clinical trials. The stakes, for patients and for the field, could not be higher.
WebProNews is an iEntry Publication