PERG is a non-invasive test of macular and retinal ganglion cell function, but its transient waveform has low amplitude and varies substantially between participants. The study examined whether diagnostic information was better retained in the complete waveform, in a smaller set of engineered numerical descriptions, or in a combined representation. Investigators evaluated three approaches under one protocol: a raw 255-point waveform that preserved amplitude; 60 descriptors covering clinical, spectral, wavelet and complexity characteristics; and a hybrid representation of 343 variables that combined both approaches and included multi-wavelet features based on the db4, sym2 and coif3 wavelets.

The dataset comprised 672 recordings from 336 examinations of 304 participants, including recordings from both eyes and repeat visits. Each recording underwent six stages of preprocessing. Model performance was estimated with 10 × 5-fold cross-validation grouped by participant, so both eyes and all repeat examinations from an individual remained in the same partition. Hyperparameter selection was nested, while every preprocessing step derived from the data was fitted within the training fold. This design reduced the possibility that repeated measurements from the same person would inflate the reported performance, although participant-grouped cross-validation cannot substitute for external, multi-centre validation.

Under the pre-specified primary classifier, the raw waveform achieved the highest pooled out-of-fold area under the receiver operating characteristic curve, or AUC, in four of the six disease-versus-control tasks. Its AUC was 0.906 for cone-rod dystrophy, 0.842 for retinitis pigmentosa, 0.755 for macular dystrophy and 0.673 for hereditary optic atrophy. The engineered descriptors ranked first in the other two tasks, with an AUC of 0.803 for Stargardt disease and 0.657 for Birdshot chorioretinopathy. The hybrid representation did not rank first in any task. It fell below the raw waveform in five of six conditions by 0.002 to 0.018 AUC, while no non-linear classifier showed a hybrid advantage above 0.006 in mean AUC. The comparison therefore did not support the idea that adding engineered features to the full waveform necessarily improved classification.

A sensitivity analysis showed why preserving amplitude mattered. When each PERG signal was z-scored to remove its absolute amplitude scale, the raw representation lost an average of 0.055 AUC, compared with an average loss of 0.023 for the engineered descriptors. This change reversed the ordering of the three representations, indicating that amplitude-related information contributed substantially to the strongest results. At the same time, reducing the engineered descriptors to 60 variables made the classification problem more tractable for linear models, which gained 0.128 in mean AUC through the dimensionality reduction. The engineered descriptors thus offered a benefit for simpler model classes, but they did not improve on the raw waveform overall.

The study’s central conclusion is specific to the six conditions and the analysed dataset: the complete amplitude-preserving PERG waveform was more effective than either replacing it with engineered features or adding those features in a larger hybrid representation. The AUC values describe model discrimination under the study’s cross-validation protocol, not diagnostic performance in routine care. The lower results for hereditary optic atrophy and Birdshot chorioretinopathy also indicate that the task remained difficult for some conditions. Participant-independent, multi-centre validation is required before the approach can be considered for clinical application, but within this internal comparison, retaining the absolute amplitude of the raw signal was central to performance.

Subject of Research: Machine-learning comparison of raw, engineered and hybrid PERG representations for classifying six retinal and optic nerve conditions

Article Title: Comparison of raw-signal, feature-based and hybrid representations for machine-learning classification of retinal diseases from pattern electroretinography

Article References: Aygün, A. B., Açıkgöz, R., Coşkun, M., & Erden, M. A. (2026). Comparison of raw-signal, feature-based and hybrid representations for machine-learning classification of retinal diseases from pattern electroretinography. Medical & Biological Engineering & Computing. https://doi.org/10.1007/s11517-026-03698-7

Image Credits: AI Generated

DOI: 10.1007/s11517-026-03698-7

Keywords: pattern electroretinography, retinal disease classification, machine learning, raw waveform analysis, feature engineering, amplitude preservation, cone-rod dystrophy, retinitis pigmentosa, Stargardt disease, Birdshot chorioretinopathy, Comparison, raw-signal

The distinction between cone-rod dystrophy and retinitis pigmentosa is not merely semantic, and the authors’ argument that CERKL disease should be regarded primarily as a cone-rod dystrophy has practical consequences for counseling and surveillance. In typical rod-cone degeneration, patients usually notice difficulty with night vision and loss of the peripheral visual field first, while central reading vision is often preserved until later in the disease course. Cone-rod dystrophy reverses this sequence: the cone photoreceptors, which are concentrated in the macula and mediate daylight acuity, color discrimination, and fine spatial detail, deteriorate earliest. This explains why the patients in this cohort presented with reduced central vision, photophobia, and dyschromatopsia, and why best corrected visual acuity at the first visit spanned such a wide range, from near-normal 20/25 vision down to light perception. For a clinician evaluating a child or young adult with unexplained central vision loss, photophobia, and a macular pattern of retinal changes, CERKL should now sit higher on the differential diagnosis list than the OMIM designation of retinitis pigmentosa alone might suggest.

The quantitative relationship between disease duration and visual acuity deserves particular attention. When the authors restricted the analysis to patients carrying two null alleles, they found a statistically robust correlation between how long the disease had been present and how much vision had been lost, with coefficients of determination indicating that a substantial fraction of the variability in acuity could be explained by disease duration alone. This kind of predictable, near-linear decline is clinically valuable for two reasons. First, it allows ophthalmologists to give patients realistic expectations about the trajectory of their condition, which is one of the most common questions in a genetic counseling session. Second, and arguably more importantly for the coming years of translational research, a measurable and consistent rate of deterioration provides the natural-history benchmark against which any future intervention would be judged. Gene therapy trials for inherited retinal disease depend on demonstrating that the treated retina declines more slowly than the untreated baseline, and cohorts like this one, with quantified progression rates, become reference points for trial design, sample size estimation, and endpoint selection.

The molecular biology of the CERKL protein remains only partially resolved, and this uncertainty shapes how the clinical findings should be interpreted. The gene encodes a 558-amino-acid protein with a diacylglycerol kinase-like domain and a predicted pleckstrin homology region, motifs typically associated with lipid signaling and membrane association. Despite the strong sequence similarity to ceramide kinase, the protein has not been shown to possess catalytic kinase activity, which has led investigators to propose that CERKL may function as a scaffolding or regulatory component of the sphingolipid pathway rather than as an enzyme. Sphingolipid metabolism occupies a central position in photoreceptor survival: ceramide is a potent pro-apoptotic second messenger, and its accumulation has been demonstrated in models of light-induced retinal degeneration, inherited dystrophies, and age-related macular degeneration. Recent work using a CRISPR-based knockdown model showed that cells lacking CERKL fail to mount their usual protective response to light stress, positioning the protein as part of the retina’s resilience machinery against oxidative damage. Photoreceptors are among the most metabolically active cells in the body, with dense mitochondrial content and chronic exposure to photo-oxidative stress, so even modest impairments in stress-response pathways can accumulate into progressive degeneration over years.

Viewed through this lens, the genotype-phenotype pattern in the cohort becomes more coherent. The dominant allele, a nonsense variant in exon 6, would be expected to trigger nonsense-mediated messenger RNA decay or produce a truncated protein lacking essential functional domains, effectively abolishing gene function. The observation that patients with two null alleles showed a consistent and predictable rate of vision loss supports a model in which complete loss of CERKL function produces a stereotyped degenerative process, presumably through the loss of oxidative stress protection in cones first and rods second. The two patients with macular dystrophy rather than full cone-rod dystrophy illustrate that residual genetic or environmental modifiers can shift the anatomical pattern of degeneration, a phenomenon seen across many recessive retinal disease genes in which the same gene can produce a spectrum from diffuse photoreceptor loss to regionally restricted macular involvement.

The methodological approach to variant classification also reflects the current standards of clinical molecular genetics. All variants were assessed under the American College of Medical Genetics and Association for Molecular Pathology framework, which integrates population frequency data from gnomAD, computational predictions of pathogenicity, and the predicted consequence of the variant on protein structure. Novel variants were accepted as pathogenic or likely pathogenic only when they produced obvious loss-of-function consequences such as frameshifts, nonsense changes, canonical splice-site disruptions, or copy-number losses, or when supporting evidence from frequency and prediction tools was sufficient. The use of two independent interpretation platforms, Franklin and Varsome, to cross-check classifications adds a layer of internal consistency, and comparison against the ClinVar archive situates the cohort’s findings within the broader community record. This rigor matters because recessive disease requires both alleles to be implicated, and misclassifying a single variant can propagate erroneous genotype-phenotype conclusions through the literature.

The epidemiological context of inherited retinal dystrophies in Latin America remains underdeveloped relative to Europe and North America, and this study contributes population-level data from a country whose ancestry is a complex mixture of Indigenous, European, African, and East Asian components. Allele frequencies established in one population cannot simply be transposed to another, and the observation that a specific nonsense allele reached homozygosity in nearly half of the affected individuals here, while other variants dominate in Spanish, Israeli, or North American cohorts, underscores how strongly founder effects and demographic history shape the genetic epidemiology of recessive disease. For genetic testing laboratories serving Brazilian patients, knowing which variants are locally enriched can improve the design of targeted panels, reduce the number of variants of uncertain significance reported, and shorten the diagnostic odyssey that families typically endure.

The diagnostic odyssey itself is worth considering. Patients with inherited retinal dystrophy often see multiple ophthalmologists over many years before a molecular diagnosis is established, and the mean delay between symptom onset and genetic confirmation remains substantial in most health systems. The clinical heterogeneity of CERKL disease, with onset anywhere from seven years of age to forty, means that affected individuals may initially receive labels ranging from Stargardt-like macular dystrophy to occult retinitis pigmentosa, depending on which examination findings dominate at the time. The wide age range at presentation also implies that disease onset is not tightly coupled to a single genotype category in this cohort, or that ascertainment bias, in which patients with earlier and more severe symptoms reach specialized clinics sooner, influences the observed distribution. Longitudinal follow-up of the cohort will help distinguish between these possibilities.

From a therapeutic standpoint, the identification of a single recurrent loss-of-function allele in a large fraction of patients is encouraging for the development of allele-specific approaches, but the broader goal in the inherited retinal disease field is genotype-agnostic intervention. The retina’s accessibility, its relative immune privilege, and the availability of non-invasive outcome measures such as spectral-domain optical coherence tomography, fundus autofluorescence, and full-field stimulus testing have made it a leading target for gene replacement and gene-editing strategies. For recessive loss-of-function genes like CERKL, the theoretical therapeutic approach is supplementation of the missing protein product, analogous to the strategy used in approved retinal gene therapy for biallelic RPE65 mutations. The predictable progression documented in this cohort would allow investigators to identify the optimal therapeutic window, likely in younger patients with preserved photoreceptor structure, before irreversible architectural loss has occurred.

The correlation between disease duration and visual acuity also raises questions about the underlying cellular dynamics. A roughly linear decline in function over time is compatible with a model in which a constant fraction of remaining photoreceptors is lost per unit of time, rather than a threshold effect in which degeneration accelerates once a critical mass of cells has been lost. Distinguishing between these models requires longitudinal structural data, ideally optical coherence tomography measurements of outer nuclear layer thickness and ellipsoid zone integrity, correlated with functional measures. The retrospective design of this study, drawing on records accumulated since 2006, means that some patients will have serial data available, and future analyses could combine structural and functional progression to refine the natural-history model further.

Finally, the study highlights the value of specialized referral centers in assembling cohorts large enough to draw genotype-phenotype conclusions for rare genes. With an estimated contribution of four to seven percent of autosomal-recessive inherited retinal dystrophy cases, CERKL is among the more commonly implicated genes in this category, yet individual centers would rarely encounter more than a handful of affected patients per year. By pooling nearly two decades of records, the investigators were able to characterize twenty-four distinct variants and establish a phenotype profile that is now better defined than it was before. As panel-based next-generation sequencing and whole-exome sequencing become more widely available and more affordable, the number of identified patients will grow, and the clinical picture assembled here will serve as the reference against which newly diagnosed individuals are compared, both for prognostication and for eligibility in future therapeutic trials.

Subject of Research: CERKL-related inherited retinal dystrophy in a Brazilian cohort: genotype-phenotype correlation

Article Title: CERKL-related inherited retinal dystrophy in a Brazilian cohort: genotype-phenotype correlation

Article References: Yasaki, E. S., Salles, M. V., Motta, F. L., & Sallum, J. M. F. (2026). CERKL-related inherited retinal dystrophy in a Brazilian cohort: genotype-phenotype correlation. Experimental Biology and Medicine, 251, Article 10935. https://doi.org/10.3389/ebm.2026.10935

Image Credits: AI Generated

DOI: 10.3389/ebm.2026.10935

Keywords: CERKL, inherited retinal dystrophy, cone-rod dystrophy, genotype-phenotype correlation, retinitis pigmentosa, Brazilian cohort, genetics, vision loss, macular dystrophy, gene variants, CERKL-related, inherited

SCA7 belongs to a family of neurological diseases caused by expanded tracts of the three-letter DNA sequence CAG embedded within a gene. In this case the culprit is the ATXN7 gene on chromosome 3, which normally encodes a subunit of a protein complex called SAGA that regulates the transcription of thousands of other genes. When the CAG tract expands beyond a pathogenic threshold, typically more than 36 repeats, the resulting ataxin-7 protein carries an abnormally long polyglutamine segment. This stretched segment drives the protein to misfold and aggregate, poisoning neurons in the cerebellum, brainstem and, uniquely among the polyglutamine ataxias, the cone and rod photoreceptors of the retina. The landmark cloning of the SCA7 gene in 1997 revealed that the CAG expansion in this disorder is among the most unstable in the human genome, and that instability underlies a notorious genetic phenomenon called anticipation.

Anticipation means that each successive generation can inherit a longer repeat tract than the parent who transmitted it, and because longer repeats generally correlate with earlier onset and more severe disease, children of affected parents may fall ill decades before their parents did. Paternal transmission is classically associated with the most dramatic expansions, and maternally transmitted cases with massive somatic instability have produced infantile-onset disease. Yet repeat length is not the whole story. Studies in transgenic mice have shown that the genomic context surrounding the repeat strongly influences its instability, and clinical experience has repeatedly shown that patients carrying identical repeat lengths can manifest very different symptoms. The new report from Zhengzhou adds compelling human evidence to that murine and statistical picture, documenting how one family under a single roof can display markedly divergent phenotypes from one shared mutation.

The centerpiece of the report is an individual who, in addition to the expected cerebellar ataxia and retinal degeneration characteristic of SCA7, exhibited angioid streaks on fundus examination. Angioid streaks appear as irregular gray or dark red lines radiating from the optic disc, visible because calcified and fragmented Bruch’s membrane allows the underlying choroidal vasculature to show through. They are classically associated with pseudoxanthoma elasticum, Paget’s disease of bone, sickle cell hemoglobinopathies and certain rare metabolic conditions, but they have never been considered a feature of SCA7. The authors documented the streaks with multimodal ophthalmic imaging: color fundus photography captured the radiating lines, fundus autofluorescence revealed corresponding linear hypoautofluorescent changes, and optical coherence tomography showed focal irregularities at the junction of the retinal pigment epithelium and Bruch’s membrane. The convergence of findings across these independent modalities strengthens the case that the structural abnormality was real and not an imaging artifact.

The clinical significance of this observation is twofold. First, angioid streaks are clinically dangerous in their own right because the brittle Bruch’s membrane predisposes patients to choroidal neovascularization, the abnormal ingrowth of fragile blood vessels that can hemorrhage and cause abrupt, severe central vision loss. A recent comprehensive review of angioid streak-related choroidal neovascularization emphasized that early detection and prompt anti-VEGF treatment can preserve vision in these patients. If SCA7 patients can develop angioid streaks, ophthalmologists caring for them may need to add this complication to their surveillance list, distinguishing new vessel growth from the slower cone-rod retinal dystrophy that ordinarily accompanies the disease. Second, from a mechanistic standpoint, the finding raises the question of whether mutant ataxin-7, which is expressed outside the nervous system, might directly or indirectly compromise Bruch’s membrane integrity, or whether the association is coincidental and merely co-occurring in this individual. The authors are appropriately cautious, presenting the case as an expansion of the phenotypic spectrum rather than proof of causation.

Equally instructive is the intrafamilial variability the report documents. A second affected individual in the same kindred showed advanced retinal degeneration with no identifiable angioid streaks at all: fundus photographs revealed diffuse degenerative changes, autofluorescence demonstrated widespread hypoautofluorescence, and OCT confirmed disruption and thinning of the outer retinal layers and retinal pigment epithelium. Two relatives, one mutation, one family environment, and yet two visibly distinct retinal phenotypes. This kind of divergence within a single pedigree is a vivid demonstration that genotype alone does not determine phenotype. Candidate modifiers include the exact repeat length and its somatic instability in different tissues, other genetic variants inherited independently at modifier loci, epigenetic differences, sex, environmental exposures and stochastic developmental events. Longitudinal cohort studies of related ataxias such as SCA1, SCA2, SCA3 and SCA6 have similarly shown that progression rates vary widely even after accounting for repeat length, and recent work on SCA1 has even uncovered sex-specific differences in presentation and progression, reinforcing the growing recognition that hereditary ataxias are shaped by networks of modifiers rather than by a single number in a sequencing report.

For clinicians, the practical message is that SCA7 should be considered a moving target. The disease has already been reported masquerading as Stargardt’s disease, an inherited juvenile macular degeneration, in patients who presented with vision loss before ataxia appeared. Some patients receive their first diagnosis from an ophthalmologist rather than a neurologist because the retina fails before balance does. Adding angioid streaks to the differential expands the diagnostic considerations in both directions: neurologists evaluating a patient with unexplained ataxia who happens to have angioid streaks on fundoscopy may now reasonably order ATXN7 repeat testing, and ophthalmologists encountering angioid streaks of uncertain cause should be alert to subtle cerebellar signs. Genetic testing remains the only definitive diagnostic tool, and the report underscores the value of combining neurological assessment, genetic confirmation and multimodal ophthalmic imaging in the workup of suspected hereditary ataxias.

The molecular biology underlying SCA7 continues to be an active therapeutic frontier. Because the disease is caused by a toxic gain of function from the expanded protein, strategies under investigation include antisense oligonucleotides designed to reduce ataxin-7 production, RNA interference approaches, gene editing to trim the repeat tract, and small molecules aimed at stabilizing the transcriptional machinery that mutant ataxin-7 disrupts. Understanding the full range of tissues affected matters for these efforts. If retinal structures beyond photoreceptors, such as Bruch’s membrane and the retinal pigment epithelium, can be drawn into the disease process, then delivery of any future therapy must reach not only cerebellar neurons but also the posterior segment of the eye, a compartment with its own pharmacological barriers. Detailed phenotyping of rare cases, as performed here, helps define the true therapeutic target list.

The study itself was a modest but careful piece of clinical science. Conducted at a single center in Zhengzhou with funding from Henan provincial research programs, it involved detailed collection and integration of clinical, genetic and imaging data, with all participants or their legal guardians providing written informed consent and the protocol approved under the Declaration of Helsinki. The authors, Chunyan Cao, Shang Xu, Yaqiong Li and Jiewen Zhang, conceived the study, assembled the data and interpreted the findings collaboratively, and they acknowledge that the observations require replication in larger cohorts before the association can be considered established. Data supporting the findings are available from the corresponding author upon reasonable request.

Rare observations like these are often dismissed as curiosities, yet history shows that they can redefine how diseases are understood and detected. Every expanded phenotypic spectrum shortens the diagnostic odyssey for someone, because a physician who has read that SCA7 can produce angioid streaks will think to test for it sooner. Every documented instance of intrafamilial variability strengthens the scientific case for hunting genetic modifiers, which in turn may reveal new therapeutic targets applicable beyond the rare families in which they are found. As gene-based therapies for polyglutamine disorders edge closer to the clinic, the precision with which the medical community can recognize every face these diseases present becomes not merely an academic exercise but a prerequisite for treating patients early enough to matter. This single family in Henan Province, examined with modern genetics and multimodal retinal imaging, has added one more facet to that recognition effort, and it is a reminder that even the best-characterized Mendelian diseases still hold surprises.

Subject of Research: Phenotypic expansion of spinocerebellar ataxia type 7, including angioid streaks and intrafamilial clinical variability

Article Title: Expanding the phenotypic spectrum of spinocerebellar ataxia type 7: angioid streaks and intrafamilial phenotypic variability

Article References: Cao, C., Xu, S., Li, Y., & Zhang, J. (2026). Expanding the phenotypic spectrum of spinocerebellar ataxia type 7: angioid streaks and intrafamilial phenotypic variability. Journal of Neurology, 273(10), Article 589. https://doi.org/10.1007/s00415-026-14081-9

Image Credits: AI Generated

DOI: 10.1007/s00415-026-14081-9

Keywords: spinocerebellar ataxia type 7, ATXN7, CAG repeat expansion, angioid streaks, phenotypic variability, retinal degeneration, cone-rod dystrophy, Bruch's membrane, genetic anticipation, hereditary ataxia, multimodal retinal imaging, choroidal neovascularization