June 2026 Research Update

Across these papers, most evidence came from single-patient reports or very small case series, so it cannot tell how often a treatment will work. Still, a consistent pattern was that some people with ongoing symptoms after early treatment improved when care added targeted immune therapies (often those that reduce specific immune cells or antibodies) and, for some, additional seizure medicines. One larger randomized trial is planned to test whether a step-by-step immune approach improves recovery in severe cases, but results were not yet available. [1] [2] [3] [4] [5] [6] [7] [8]

For anti-NMDA receptor encephalitis that happens with ovarian teratomas, two single-patient reports described improvement after combining immunotherapy with surgery to remove the tumors, alongside prolonged hospital-level supportive care. One case included a long hospital stay with complications that were managed during multidisciplinary care. [2] [5]

For anti-NMDA receptor encephalitis that continued despite multiple earlier treatments, a three-patient series reported rapid clinical improvement after subcutaneous ofatumumab, along with marked lowering of CD20-positive B cells, and only mild short-term side effects in one patient. No relapses were reported during follow-up (15 months to 2 years), but the lack of a control group makes it hard to know how much improvement was due to the drug versus other factors. [3]

A multicenter open-label randomized trial in China (RADIA) is designed to test whether sequential ofatumumab followed by daratumumab improves functional outcomes in severe autoimmune encephalitis (mostly NMDA receptor encephalitis) compared with ofatumumab alone or repeated intravenous immune globulin or plasmapheresis. Because this paper is a study protocol, it does not yet provide results on benefits or risks. [4]

Several reports focused on specific antibody types and individualized responses. In one person with LGI1 antibody limbic encephalitis and frequent brief arm-and-face spasms, symptoms improved with steroids and a seizure medicine (lacosamide), and later resolved after rituximab, while plasmapheresis was tried without benefit in that case. [1]

In autoimmune GFAP astrocytopathy, one teen who did not improve with steroids, intravenous immune globulin, or rituximab had marked and sustained improvement after five infusions of tocilizumab, with brain scan and EEG changes also improving; the paper also identified one prior similar case in the literature. This suggests IL-6 blocking treatment may help some refractory cases, but evidence remains very limited. [6]

For anti-KLHL11 encephalitis, two adults had different responses: one recovered after intravenous immune globulin plus low-dose rituximab, while the other relapsed after standard immunotherapy and improved after off-label efgartigimod; one patient also had medication-induced movement side effects from haloperidol. The authors emphasized clinical variability and noted the need for ongoing evaluation for an underlying tumor in this antibody type. [7]

For seizure control in children with autoimmune encephalitis, a retrospective 14-patient series reported that some improved after adding perampanel, with no serious side effects reported, although mild behavioral and dizziness-type side effects occurred. Because many children also received immunotherapy and multiple seizure medicines, it was not possible to clearly separate perampanel's role from other treatments. [8]

For patients and caregivers, these papers mainly add context for what clinicians may try when symptoms continue after early treatment, especially in specific antibody-defined types (LGI1, NMDA receptor, GFAP, KLHL11). The strongest information in this set is still limited because most reports describe one person or a few people, so benefits and side effects could look different in larger groups. The planned randomized trial matters for the future because it aims to test, in a more reliable way, whether stepwise immune treatments improve functional recovery in severe disease; until results are published, uncertainty remains about how well this approach works and for whom. [4]

This month included one detailed rehabilitation case report rather than a larger study, so it cannot show what is typical. Even so, it illustrates that meaningful recovery in movement, daily activities, thinking, and language can occur with structured therapy, even when rehabilitation begins long after the initial illness. The report did not include comparisons, so it is uncertain how much improvement was due to rehabilitation versus delayed recovery from the disease itself. [9]

In a 35-year-old woman with anti-NMDA receptor encephalitis, intensive rehabilitation in the intensive care unit started about 20 months after onset and ran for about 3 months (physical, occupational, and speech therapy, five sessions per week). The patient improved from complete dependence to substantial independence in daily activities, regained independent walking, and showed measurable gains in balance, hand function, thinking, and language testing. No rehabilitation-related adverse events were reported during continuous monitoring. [9]

For patients and caregivers, this report is most relevant to those dealing with long-lasting disability after anti-NMDA receptor encephalitis, including severe weakness, difficulty with speech, or dependence on hospital-level support. It suggests that formal, goal-directed rehabilitation may still coincide with major functional gains even late in the course, but because it is one person, it does not answer how often this happens or what intensity and setting are best. More research is needed to compare different rehabilitation approaches and to understand which symptoms (movement, thinking, speech) are most likely to improve and on what timeline. [9]

Across these studies, predictors of recovery and relapse fell into a few practical categories: how severe the illness is early on, which antibody type (if any) is involved, and certain test results (brain MRI changes, blood markers, and antibody levels). Several studies agreed that people who were sicker during the acute phase were more likely to have worse function months later, while longer-term relapse risk was reported most clearly in antibody-negative (seronegative) groups. [10] [11] [12]

In a 150-person cohort, poor function at 6 months was more likely in people who needed mechanical ventilation, had a tumor, had very impaired consciousness, and were older. About 30% had a poor outcome by that definition, highlighting that early critical illness features can signal a harder recovery. [10]

For LGI1 encephalitis, one multicenter MRI study found that acute hippocampal swelling (a temporary enlargement of a memory-related brain structure) was linked to worse function at 12 months and to later scarring-like changes on follow-up scans. In that same study, a bright signal on MRI by itself was generally not enough to predict 12-month function, even though it related to later tissue loss in LGI1 cases. [13]

A small LGI1 cohort suggested that higher LGI1 antibody levels in both blood and spinal fluid, older age, and longer delay to diagnosis were tied to worse 12-month function, even though many people had normal-looking early MRI or spinal fluid routine tests. This was an exploratory study, so it suggests possible warning signs rather than giving definitive cutoffs for individual prediction. [14]

In anti-NMDAR encephalitis, a 186-person study reported that a higher neutrophil-to-HDL ratio (a blood test combining an inflammation-related cell count with "good" cholesterol) predicted greater severity and worse short-term status at hospital discharge. This points to measurable inflammation in the blood as one way researchers are trying to estimate near-term prognosis, though it does not address long-term relapse. [15]

For people with seronegative autoimmune encephalitis (no identified neuronal antibody), a meta-analysis found that about two-thirds reached good functional status overall, with better recovery in seronegative limbic encephalitis than in antibody-negative probable autoimmune encephalitis. The same review estimated relapse at about 17% and reported a higher risk of ongoing seizures in seronegative groups compared with seropositive groups, while cognitive outcomes were reported inconsistently across studies. [11]

In children with probable antibody-negative autoimmune encephalitis, most had a favorable long-term outcome in one cohort, but relapses still happened in a minority and occurred around the first year after onset. In that study, all children improved after restarting immune treatment, but this does not prove what will happen in other settings or age groups. [12]

Even when physical function improves, longer-term emotional and behavioral symptoms were common across encephalitis survivors in a large meta-analysis, including in autoimmune causes such as anti-NMDAR and anti-LGI1. The review reported substantial differences between studies and measurement tools, so exact rates vary, but it underscores that "recovery" can involve mental health and behavior, not only movement and self-care. [16]

One study focused on recovery in a specific life domain: reproductive hormones and fertility in women with anti-NMDAR encephalitis. It found that ovarian hormone suppression was common during the acute illness, tended to improve during recovery, and pregnancy rates were similar in teratoma-associated and non-teratoma cases, even though teratoma-associated illness was more severe early on. [17]

For patients and caregivers, this research helps explain why recovery can look different from one person to another: early illness severity (like needing a ventilator or having severely reduced consciousness), certain MRI changes in LGI1 encephalitis (especially hippocampal swelling), and some lab measures (LGI1 antibody levels or an inflammation-related blood ratio in anti-NMDAR) were linked to worse short-term or medium-term outcomes in observational studies. Relapse was reported as a real possibility in antibody-negative groups, and persistent seizures and longer-term mood or behavior changes were repeatedly described as part of the post-encephalitis experience. What remains uncertain is how well any single test can predict an individual person's outcome, because many results came from retrospective studies with different follow-up timing and varying outcome measures. [10] [13] [14] [15] [11] [12] [16] [17]

Across these papers, researchers explored diagnosis-supporting signals that could add context when autoimmune encephalitis is suspected, especially when antibody tests are negative or still pending. The overall pattern was that several tests showed measurable differences in groups with autoimmune encephalitis, but the same markers could also rise in other inflammatory brain conditions, so they did not cleanly separate autoimmune encephalitis from all alternatives. Most evidence here is exploratory and based on small groups, so it is best read as early work on additional pieces of the puzzle rather than stand-alone answers. [18] [19] [20] [21]

In a cross-sectional study, multiple immune signaling proteins measured in cerebrospinal fluid (spinal fluid) were higher in people with autoimmune encephalitis than in controls, and the biggest differences were seen for CXCL10 and CXCL13. Importantly, people labeled "seronegative" (no known antibodies found) had similar spinal fluid immune patterns to people with known autoimmune encephalitis antibodies, suggesting these immune markers might act as supportive information while antibody results are being clarified. The authors emphasized the work was exploratory and noted limits such as the control group being mostly people with idiopathic intracranial hypertension and few infection comparisons. [18]

A separate study focusing only on spinal fluid CXCL13 across several inflammatory brain conditions found CXCL13 was highest in Lyme neuroborreliosis and did not show a disease-specific pattern for viral encephalitis or autoimmune encephalitis. CXCL13 levels also tracked with more general spinal fluid inflammation measures like cell count and protein, which helps explain why the same marker can rise in multiple diagnoses. This highlights a key uncertainty: a high immune marker may reflect "inflammation is happening" without pinpointing the exact cause. [19]

In children with seronegative autoimmune encephalitis, an MRI-based automated brain volume approach found smaller grey matter and certain deep brain structure volumes than healthy controls, with putamen and pallidum differences still present after adjusting for head size. The study also reported that some volume measures could distinguish cases from controls in statistical analyses, but it was small and cross-sectional, so it cannot yet show how reliable this would be for real-world diagnosis across different hospitals and scanners. The main takeaway is that quantitative MRI measurements may add information beyond a standard visual read, but need more validation. [20]

In adults with definite N-methyl-D-aspartate receptor (NMDAR) antibody encephalitis, seizures were common at presentation and electroencephalography (EEG, a brain-wave test) often showed diffuse slowing with variable epileptiform activity. A specific pattern called extreme delta brush was rare, and in one case it appeared alongside non-convulsive status epilepticus. This study describes what clinicians may see on EEG early in NMDAR encephalitis, but it does not establish that these EEG changes are unique to this condition. [21]

For patients and caregivers, these studies help explain why diagnosing autoimmune encephalitis often uses multiple types of information: symptoms and exam, antibody testing, spinal fluid results, brain imaging, and sometimes EEG. This month’s papers suggest several add-on measures (spinal fluid cytokines like CXCL13 or CXCL10, quantitative MRI volumes, and EEG patterns in NMDAR encephalitis) can provide supportive clues, but they can overlap with other inflammatory conditions and still need larger, confirmatory studies. What remains uncertain is how well these measures perform in everyday practice across diverse patients, especially when trying to separate autoimmune encephalitis from infections or other causes of brain inflammation. [18] [19] [20] [21]

Two studies in people evaluated for memory problems showed that simply finding antibodies does not always equate to autoimmune encephalitis. In these settings, likely autoimmune encephalitis was rare, and many antibody-positive results did not clearly fit an autoimmune encephalitis syndrome, raising the risk of confusing or incidental findings when screening is done without strong clinical suspicion. These papers reinforce that antibody results are best interpreted alongside the full clinical picture rather than treated as a stand-alone answer. [22] [23]

In a retrospective screening study of 1021 memory-clinic visitors, only 11 people were antibody-positive on repeated testing, and just three were judged to have phenotypes suggesting autoimmune or paraneoplastic encephalitis. The authors concluded that unselective screening in memory clinics is not suitable because the frequency of likely autoimmune encephalitis was low. Brain MRI in these patients did not show active inflammation in the medial temporal lobes, and other findings such as atrophy or small-vessel disease were variable. [22]

In a cross-sectional study of 97 people with mild cognitive impairment, antibodies were detected in 10 patients across several targets (including anti-GABA B receptor and several intracellular antibodies). The antibody-positive and antibody-negative groups looked clinically similar overall, while tumor markers were more often abnormal in the antibody-positive group. In a small subset of amyloid-positive patients who had brain imaging, the antibody-positive group showed relatively preserved brain metabolism on FDG-PET, while amyloid PET did not differ between groups. [23]

For patients being evaluated for memory decline, these studies mainly serve as background on the limits of broad antibody testing: a positive antibody result can occur without a clear autoimmune encephalitis presentation, and may not by itself explain symptoms. These results matter now because they highlight uncertainty in how to interpret antibody tests in slow-onset cognitive settings, where other causes of symptoms are common. What remains uncertain is which combinations of symptoms, test results, and antibody types best identify the smaller group of people whose problems are truly driven by immune attack on the brain. [22] [23]

These papers show that autoimmune encephalitis can be hard to recognize because its early symptoms can look like other brain and mental health conditions. In several reports, people first appeared to have a primary psychiatric disorder, a rapidly progressive dementia, or another inflammatory brain syndrome, and in one case a brain tumor created a picture that resembled autoimmune encephalitis. [24] [25] [26] [27]

One challenge was that blood, spinal fluid, MRI, and EEG tests were not always clearly supportive of autoimmune encephalitis early on. Some patients had no identifiable autoimmune encephalitis antibodies (sometimes called seronegative disease), while others had antibodies that were not on standard panels or were hard to interpret. [28] [29] [27]

Several cases highlighted how autoimmune encephalitis can resemble neurodegenerative disease, including Alzheimer-like symptoms and even Alzheimer-like blood marker patterns. In one patient, anti-mGluR5 autoimmune encephalitis was initially labeled as rapidly progressive Alzheimer disease because the clinical picture, imaging, and blood markers seemed to fit, before antibody testing supported an autoimmune cause. [25]

Another key mimic described here was brain tumor: a 59-year-old man had seizures and a low-level CASPR2 antibody in blood, which initially suggested autoimmune encephalitis, but later imaging and biopsy confirmed glioblastoma. After tumor surgery, the antibody became negative and his seizures did not recur during follow-up, supporting the idea that some tumors may trigger immune signals that look like autoimmune encephalitis. [27]

Thyroid-related steroid-responsive encephalopathy (sometimes called SREAT) was also presented as a look-alike, with acute psychosis and neuropsychiatric decline plus high anti-thyroid antibodies and subclinical hypothyroidism. That case had rapid improvement in mental status within 48 hours after high-dose steroids, after other causes were assessed. [26]

Finally, these reports stress that anti-NMDA receptor encephalitis and other autoimmune encephalitis types can occur outside the typical age groups and can initially be mistaken for a primary psychotic disorder, including in older adults. In one 59-year-old woman with subacute delirium and behavioral changes, spinal fluid antibody testing ultimately confirmed anti-NMDA receptor encephalitis after earlier evaluations did not provide answers. [24]

For patients and caregivers, this group of papers mainly serves as background on why the path to an autoimmune encephalitis diagnosis can be complex and sometimes indirect. The cases illustrate that similar symptoms (confusion, psychosis, memory loss, seizures) can come from different causes, that antibody tests can be negative or hard to interpret in some people, and that other illnesses (including thyroid-related brain inflammation or a brain tumor) may look similar early on. Because these are mostly single-patient reports, they cannot show how often any mimic occurs, and they cannot predict what will happen for an individual person.

Taken together, these papers suggest that autoimmune encephalitis can show up alongside other immune problems or after immune system disruptions such as viral infections, immune-targeting cancer therapy, biologic immune medicines, or cranial radiotherapy. Several reports describe overlap syndromes where more than one antibody pattern was present, which can blur the line between different encephalitis labels. Most evidence here is exploratory and based on case reports, with one small single-center case series, so it cannot show how common these situations are.

Two reports describe autoimmune encephalitis occurring in the setting of possible viral involvement detected in spinal fluid. One 49-year-old man had both anti-N-methyl-D-aspartate receptor (NMDAR) and anti-myelin oligodendrocyte glycoprotein (MOG) antibodies, and cytomegalovirus (CMV) genetic material was detected in spinal fluid by metagenomic sequencing; he improved after antiviral therapy plus immunotherapy, and CMV testing later became negative. [30] Another report described a 22-year-old woman whose illness appeared to shift from an initial viral encephalitis phase to a later autoimmune phase; during relapse, metagenomic sequencing detected HHV-7 and Epstein-Barr virus (EBV) in spinal fluid, and thyroid antibodies rose with relapses, with improvement after immunotherapy. [31]

Overlap between antibody-defined syndromes was a repeated theme. In a 10-person series of MOG-associated cerebral cortical encephalitis, those who also had NMDAR antibodies in spinal fluid had more psychiatric symptoms, abnormal behavior, and movement disorders, yet functional outcomes were similar between groups at follow-up. [32] A separate single-patient report described a rare triple-positive pattern (NMDAR-IgG, MOG-IgG, and mGluR5-IgG), with complete cognitive recovery within three months and antibody levels dropping over time. [33]

Several reports focused on autoimmune encephalitis appearing after immune-modifying therapies or radiation, suggesting that immune balance changes may sometimes coincide with brain inflammation. One patient developed autoimmune encephalitis as an immune-related adverse event during atezolizumab plus bevacizumab treatment for liver cancer and improved rapidly after high-dose steroids. [34] Another young woman developed autoimmune encephalitis with axonal polyradiculoneuropathy after adalimumab and secukinumab, with partial response to intravenous immunoglobulin and steroids and stronger improvement after rituximab. [35] In a patient with thymoma-associated myasthenia gravis, anti-LGI1 encephalitis developed while taking efgartigimod even though blood immunoglobulin G levels were lowered, and cognition improved after high-dose steroids. [36] One additional case linked anti-AK5 autoimmune encephalitis to the period after craniopharyngioma surgery and cranial radiotherapy, with later seizures and hippocampal changes on MRI. [37]

For patients and caregivers, this set of studies mainly provides background on situations that can overlap with autoimmune encephalitis: viral genetic signals in spinal fluid, multiple antibody patterns in the same person, and onset after treatments that strongly affect the immune system (including some cancer immunotherapies, biologic immune drugs, and possibly brain radiotherapy). Because most reports describe single individuals, it remains uncertain how often these overlaps happen, whether they are true triggers versus coincidental timing, and which test results best separate infection, autoimmune disease, and mixed pictures in real-world care.

Across this month’s papers, autoimmune encephalitis showed wide variation by antibody type, but several recurring symptom patterns stood out: (1) limbic symptoms such as new memory problems, behavior changes, sleep disruption, and seizures, (2) prominent balance and coordination problems (ataxia), and (3) in some antibodies, strong body-wide symptoms such as severe abdominal pain. A common practical point across multiple antibodies was that standard brain MRI could be normal or only mildly abnormal, and antibody testing was often more likely to be positive in blood than in spinal fluid for some syndromes. [38] [39] [40] [41]

For anti-mGluR5 encephalitis, two datasets reinforced that many people present with cognitive and behavioral symptoms, sleep problems, and seizures, and about half may have a normal brain MRI. Both reports also noted that the antibody was much more often detected in blood than in spinal fluid, meaning spinal fluid testing alone did not capture all cases in the literature they reviewed. [38] [39]

For anti-DPPX encephalitis, a single patient and a larger literature review highlighted that severe abdominal pain can be a major early symptom along with seizures, and that blood testing was more often positive than spinal fluid testing in reported cases. The review also described relapses as relatively common across published cases, showing that symptom return can be part of this syndrome for some people. [40]

Several papers focused on antibodies linked to prominent ataxia (balance and coordination problems), sometimes with normal MRI results. One case tied anti-TR(DNER) antibodies to a paraneoplastic syndrome with lymphoma and reported stabilization after combined immune and cancer treatment, and another case described anti-Ma2 autoimmune cerebellar ataxia (a related immune condition of the cerebellum, the brain's balance center, rather than encephalitis) initially resembling Hashimoto-related disease, with improvement after intravenous immunoglobulin and improved cerebellar blood-flow imaging. [41] [42]

Anti-GAD65 (glutamic acid decarboxylase 65) antibody-associated disease was presented as clinically mixed, often involving seizures and sometimes other immune conditions, with antibody levels not always matching day-to-day symptoms in a small adult series. A separate small pediatric series emphasized that GAD65 antibodies can appear in children with autoimmune brain conditions, adding to limited child-focused reports. [43] [44]

Not all relevant immune-related neurologic syndromes fit neatly into a single-antibody, single-trigger pattern. One report described a person with both anti-NMDAR and anti-MOG antibodies and cytomegalovirus (CMV) genetic material detected in spinal fluid, improving after antiviral therapy plus immunotherapy. Another pediatric case linked recurrent seizures and encephalopathy to celiac disease markers and intestinal biopsy, with long seizure freedom reported after diet change in that single patient. [30] [45]

In children with acute ataxia (a broader group than autoimmune encephalitis), a moderate-strength retrospective case-control study found that immune-mediated cases had worse short-term disability and more relapses than seronegative controls, and it identified features that were more common in children who tested positive for myelin oligodendrocyte glycoprotein (MOG) antibodies. The MRI pattern in MOG-positive children more often involved multiple brain regions, not only the cerebellum. [46]

A few papers looked beyond bedside symptom patterns to underlying biology and additional testing tools. A laboratory study using patient-derived antibodies suggested that, in multiple autoimmune encephalitis types, some autoantibodies may already bind strongly to nerve-cell surface targets even before the usual process of antibody "fine-tuning" in the immune system, and some showed cross-reaction with other self-targets in lab tests. A mouse model of anti-AMPAR encephalitis showed progressive neurologic symptoms with antibody deposits and reduced receptor signals in the brain, supporting the idea that ongoing antibody activity inside the central nervous system can drive disease features over time. [47] [48]

Two smaller clinical reports highlighted that imaging and body systems outside the brain can matter when sorting syndromes. A two-patient anti-AK5 series described prominent new memory formation problems plus psychiatric and sleep symptoms, with a specialized inflammation-sensitive brain scan showing increased signal in multiple regions, and outcomes that differed between a teenager and an older adult. Another single anti-LGI1 case described a slow, months-long cognitive decline before seizures, along with low sodium levels and characteristic MRI and brain metabolism scan changes, followed by partial cognitive recovery. [49] [50]

Finally, one case report underscored that people with autoimmune encephalitis may also experience significant non-neurologic immune problems during their illness course, such as severe psoriasis that did not resolve after stopping seizure medicines and improved only after psoriasis-directed biologic therapy. Another case showed that a different immune condition can cause similar brain symptoms: a young woman with speech and planning difficulties was diagnosed with autoimmune thyroiditis and improved after thyroid hormone replacement. Together, these single-patient reports show how other immune conditions, whether alongside autoimmune encephalitis or as a look-alike, can complicate the overall picture. [51] [52]

A small exploratory microbiome study that included an autoimmune encephalitis group found differences in gut bacteria compared with controls and other epilepsy groups, but the results were early and could be influenced by diet and antiseizure medicines. This line of work is still mainly background research rather than a guide to specific care decisions. [53]

For patients and caregivers, these papers mainly help explain why symptoms and test results can look very different across autoimmune encephalitis types, even when the condition is immune-driven. The strongest near-term takeaway is descriptive: certain antibodies tended to cluster with specific problems (for example, severe abdominal pain in anti-DPPX, or ataxia in some paraneoplastic antibodies), and normal MRI or negative spinal fluid antibody tests did not rule out these syndromes in published cases. What remains uncertain is how well these patterns predict an individual person’s course, because many reports were small, retrospective, or single-patient descriptions, and laboratory and mouse studies cannot directly tell what will happen in humans.