Current knowledge of intermittent photic stimulation for electroencephalography in dogs

Abstract

Intermittent photic stimulation is a standardized activation procedure for electroencephalography in humans that increases the detection of epileptic activity and improves diagnostic yield. However, its application in veterinary medicine remains inconsistent and methodologically heterogeneous. This narrative review synthesizes current knowledge regarding intermittent photic stimulation in dogs, critically evaluates variability in stimulus design, vigilance state, and electrode montage, and proposes a framework to support standardization in veterinary practice. Published veterinary studies demonstrate that photic stimulation can provoke generalized spike-wave and polyspike-wave discharges in selected epilepsy syndromes of dogs, particularly during awake recordings. However, protocols vary substantially in frequency range, light intensity, eye state control, flash duration, vigilance state, and lamp distance. Sedation and general anesthesia appear to attenuate retinal and thalamocortical responsiveness, decreasing the likelihood of eliciting photoparoxysmal responses. In addition, sparse and inconsistent electrode arrays limit spatial interpretation of photic responses and hinder cross-study comparisons. Drawing on established methodology used in humans and emerging veterinary data, this review outlines practical recommendations for standardized awake testing, defined stimulus parameters, and improved spatial sampling. Harmonized protocols may enhance diagnostic sensitivity, facilitate identification of photosensitive epilepsy syndromes in dogs, and strengthen translational alignment between epileptology in humans and animals. Standardization of intermittent photic stimulation represents an opportunity to improve the clinical utility and reproducibility of electroencephalography in dogs.

Introduction

Understanding what disrupts cerebral homeostasis sets the stage for how seizures are conceptualized and epilepsy is defined. Seizures arise when hyperexcitable cortical circuits briefly generate aberrant neuronal activity that recruits broader networks and produces clinical signs.¹ Epilepsy is the disease state in which a brain has more than one unprovoked seizure at least 24 h apart or one unprovoked seizure with a high risk of recurrence.² An epilepsy syndrome is a specific diagnostic construct that groups patients by shared features, such as seizure types, age at onset, electroencephalographic patterns, etiologic findings, and expected course.^3–5 People benefit from a mature systemic understanding of epilepsy semiology resulting in dozens of available diagnoses organized by the characteristics mentioned. However, veterinary characterization of epilepsy syndromes lags behind classifications available in epileptology of humans. Although their clinical manifestations and therapeutic responses closely parallel those seen in people,⁶ the systematic categorization of epilepsy syndromes in veterinary medicine remains limited to only 3 classifications. The International Veterinary Epilepsy Task Force classifies epilepsies in dogs as idiopathic, structural, or unknown, emphasizing a taxonomy that is far less granular than the syndrome-based framework developed in people.⁷ Clinical understanding of epilepsy in humans has advanced considerably through the widespread use of electroencephalography (EEG), a technique that measures brain activity via electrodes placed on the scalp. These electrodes capture epileptiform discharges (EDs) which are brief episodes of abnormal cortical patterns that appear during and between epileptic seizures. This approach has allowed EEG to become a crucial diagnostic tool for identifying and classifying epilepsy syndromes, guiding both medical and surgical treatment decisions, as well as therapeutic planning and prognosis. The spontaneous and unpredictable nature of EDs means that routine veterinary EEG will not capture them on every recording session.^8,9 As a result, EEG sessions are often supplemented with activation techniques, such as intermittent photic stimulation (IPS). Unlike waiting for spontaneous discharges, IPS offers the advantage that photosensitive EDs can be evoked under controlled conditions, enabling standardized assessment across recordings and patients, informing practical recommendations.^10,11

By exposing the patient to periodic flashes at different frequencies, IPS can provoke characteristic responses in the visual cortex and interconnected cortical–subcortical networks. Current terminology has refined the classification of IPS-induced EEG responses into 3 distinct categories: photic driving response (PDR), orbitofrontal photic myoclonia, and photoparoxysmal responses (PPRs).^12,13

Firstly, the PDR is a neurotypical physiological entrainment of occipital rhythms that mirrors a stimulus frequency. A PDR carries no epileptic relevance, may be symmetric or asymmetric across hemispheres, and fades promptly when stimulation ceases.^14,15 Secondly, orbitofrontal photic myoclonia is a nonepileptic frontal response characterized by brief muscle artifact over the orbitofrontal regions, time-locked to flash onset.^12,13 It has its own distinct category because it can be mistaken for true EDs if not carefully distinguished. In contrast, PPRs are visual-specific EDs that reflect pathological cortical hyperexcitability.^12,13 In susceptible individuals, the same afferent drive that produces benign PDRs, instead tips corticothalamic circuits into self-sustaining discharges that may persist beyond the stimulus.¹⁶ Morphologically, PPRs are categorized into a 4-grade system: occipital-limited spikes (grade 1), parieto-occipital spikes with a biphasic slow wave (grade 2), all of the features of grade 2 but also present in the frontal region (grade 3), and generalized spike- or polyspike-wave complexes (grade 4).^17–19 These grades have provided valuable diagnostic information in characterizing photosensitive epilepsy syndromes in people.^18,19 As such, the clinical utility of IPS has been embedded in routine EEG in humans since standardization of the procedure in 1999, improving diagnostic sensitivity and decreasing time to definitive diagnosis.²⁰

Overall, these systematic understandings and protocols highlight not only the importance of activation techniques for epilepsy diagnosis in people but the opportunity to extend this knowledge to other species.⁶ Although the principles of visual processing mechanisms involved in IPS are conserved across mammalian species,²¹ IPS has yet to be systematically integrated into routine EEG in dogs. This gap underscores the need to determine whether IPS can improve diagnostic yield and efficiency in EEG protocols for dogs, thereby strengthening both clinical practice and the translational value of epilepsy research in dogs. This narrative review aims to consolidate the level of current understanding of IPS protocol in dogs, determine its current utilization in routine epileptic EEGs, identify barriers in our understanding, and offer suggestions to promote its future use in routine EEG in dogs.

From light to neural response

By targeting the visual cortex, IPS challenges how the brain processes visual signals. Flashes from a stimulatory source activate retinal photoreceptors that relay signals through bipolar and retinal ganglion cells.²² The action potentials resulting from that stimulus travel down the optic nerve, decussate at the optic chiasm, and project via the optic tract to the lateral geniculate nucleus (LGN) of the thalamus.²² From the LGN, thalamocortical projections deliver the input to the primary visual cortex, where responses propagate through occipital and associated networks.²² This well-mapped retina–geniculate–cortical chain is the substrate that IPS perturbs, and the pathway through which periodic stimuli can amplify thalamocortical responses in susceptible brains.

The standard protocol for IPS in people involves generating a train of rhythmic flashes from a 30 cm distance, and at a fixed light intensity. Light intensity is set at the highest available lamp setting, because it shapes visual drive and the probability of a PPR.²³ Furthermore, lamp flicker frequencies typically are delivered in separate flash trains, with a minimum total duration of 7 s per frequency, but longer durations are preferable when 3 eye-state conditions are tested sequentially.²⁴ Frequencies are presented in an ascending order from low to high, and then descending, typically covering ranges from 1 to 60 Hz.^20,24 In humans, the recommended frequency sequence is 1, 2, 8, 10, 15, 18, 20, 25, 40, 50, and 60 Hz.^24,25 If a generalized PPR occurs at a given ascending frequency, it establishes a lower provocation threshold. The remaining ascending frequencies then are skipped and testing resumes at 60 Hz, descending through the sequence until another generalized PPR recurs, establishing the upper threshold. When doubt exists about whether a particular frequency has provoked a generalized PPR, that frequency should be repeated after a rest of 10 s, or a frequency differing by 1 Hz should be tested.²⁴ The explicit goal of this approach is to define both the lower and upper provocation thresholds for each patient, creating a quantifiable range of photosensitivity that supports both clinical decision-making and cross-study comparison. Importantly, a minimum rest period of 5 s must be observed between frequency changes, and 10 s if a generalized PPR is observed to decrease the risk of cumulative cortical excitation.^20,24 Eye state is another major determinant of eliciting PPRs during IPS.

Testing stimulation with eyes open, eyes closed, and at eye closure yields different results, with transition at eye closure often being the most provocative condition for eliciting EDs.²⁴ Empirical work shows that keeping the eyes closed during stimulation increases the likelihood and magnitude of PPRs compared with eyes open, an effect attributed in part to eyelid-mediated light diffusion and to state-dependent changes in occipital excitability.²³ Physiologically, closing the eyes enhances posterior α synchronization and removes central fixation, changing thalamocortical circuitry bias toward hypersynchrony.²⁶ Clinically in humans, eye closure sensitivity is enriched in photosensitive generalized epilepsies, most notably epilepsy with eyelid myoclonia, where eye closure and IPS reliably trigger EDs, underscoring that eye state is not a trivial parameter but a syndrome-relevant probe of cortical hyperexcitability.^27,28 For this reason, technical guidelines emphasize documenting eye condition when reporting IPS protocols.²⁹ However, applying this standard from humans to dogs is not straightforward. In practice, differentiating eye states in dogs is likely to provoke resistance and stress which may attenuate cortical responsiveness. Such practical constraints illustrate why veterinary IPS protocols require dedicated optimization rather than simple adoption of methods used in humans.

Additionally, species-specific visual physiology also shapes how dogs respond to photic stimulation. Dogs have higher critical flicker fusion thresholds than humans, with values of approximately 70-80 Hz in dogs compared with 50-60 Hz in humans.³⁰ Consequently, flash rates that appear steady to clinicians still may be perceived as flicker by dogs, an important consideration that should inform the choice of IPS frequencies if possible.³⁰ Additionally, colored flashes such as alternating chromatic patterns are sometimes utilized in IPS protocols used in humans because certain wavelengths increase the chances of PPRs.^31,32 This implementation must be considered cautiously because dogs have dichromatic vision based on 2 cone types biased toward blue and yellow.^30,33,34 Therefore, reds and many greens are poorly distinguished, making chromatic IPS paradigms optimized for people unlikely to engage visual pathways in dogs to the same degree.

The current position of IPS use in dogs

Use of IPS in EEGs in dogs falters primarily because of the lack of standardization, which has produced substantial variation and inconsistency in the published literature. The veterinary EEG literature on IPS in dogs begins with a study that first incorporated flash trains into EEG in dogs and demonstrated that IPS could provoke PPR activity in a subset of cases.³⁵ This early work provided critical proof-of-principle that light-driven activation can increase diagnostic yield beyond spontaneous EDs in dogs. Over the next 4 decades, the variability of IPS use in dogs remained sporadic, with major methodological inconsistencies such as varying sedation techniques, brief protocols, a focus on genetic breed predispositions, and narrow frequency sweeps—all of which likely suppressed IPS sensitivity and hampered cross-study comparison (Table 1).

Table 1.

Summary of variability in IPS protocols across published EEG studies in dogs.

Study	n	State	Sedative agent	Frequency range (Hz)	Train/Rest (s)	Distance (cm)	Electrode count	IPS outcomes
Holliday et al., 197035	13	AW	NR	5 - 25	30 - 40/NR	NR	6	PPR in subset
Wrzosek et al., 201736	125	SED	MED	0.5-60	NR	NR	10	No new ED during IPS
Brauer et al., 201137	10	AN	PROP	5-50	8/5	20	5	No PPR
Brauer et al., 201238	89	AN	PROP	5-50	8/5	20	5	No new ED during IPS
Demeny et al., 202039	2	SED	MED	1-16	10/5	30	12	PSW + myoclonic jerks
Wielaender et al., 201740	10	AW	NR	1-25	10/5	30	15	SW/PSW; PPR+
Wielaender et al., 201841	1	AW	NR	1-25	10/5	30	15	No PPR; absence seizures
Lyon et al., 202442	45	AW	NR	3-50	10/5	30	10	PDR + PSW

Abbreviations: AN = anesthesia; AW = awake; ED = epileptiform discharge; IPS = intermittent photic stimulation; MED = medetomidine; NR = not reported; PDR = photic driving response; PPR = photoparoxysmal response; PROP = propofol; PSW = polyspike-wave; SED = sedated; SW = spike-wave.

Variability in stimulus design

Veterinary IPS protocols differ widely in the stimulation design to elicit PPRs in 3 areas: light intensity, lamp distance, and flash train blocks. Early work reported broad frequency bands and long blocks without technical detail on light output. For example, the previously mentioned study used flash rates between 5 and 25 Hz and kept each rate on for 30-40 s, but did not report distance from the eye.³⁵ More recent veterinary studies improved structure but still varied in key factors that limit comparability. A large, sedated cohort used a continuous sweep that increased from 0.5 to 60 Hz over 5 min and then returned to the base point, but did not report eye state or the lamp distance.³⁶ Another study observing IPS in anesthetized dogs specified lamp distance and a stimulation block design.^37,38 These studies took a different approach, setting lamp distances to 20 cm away from closed eyes, and shorter flash trains such as 5-50 Hz in 5 Hz increments, and applied each rate for 8 s with 5 s of rest.^37,38 This level of detail allows others to reproduce the response pattern of generalized polyspike-slow-wave discharges time-locked after the flashes.^36,37 The stimulus design differences in distance, vigilance state, and flash-train frequency spacing complicate any attempt to pool detection rates or to analyze frequency-specific reactivity.^37–39 Defined frequency blocks and rests are a step forward, but without intensity, fixed lamp distance, and eye state differential, the biological meaning of successful IPS remains hard to compare. Moreover, reports of awake EEG that use structured IPS show further progress in clarity of outcomes. These studies documented time-locked photoconvulsive responses and generalized polyspike or spike-wave activities during ambulatory recordings with IPS, but did not provide the exact light intensity (Table 1).^40,41

In contrast, guidelines used in humans explicitly describe the order of eye states, an exact distance of 30 cm, an awake vigilance state, and the way intensities and durations are controlled and reported.^20,24 Methodology papers in humans and vision studies in dogs both showed that stimulus intensity and duration alter response probability and timing, but these parameters are usually missing from IPS reports in dogs.^20,43 Overall, the veterinary IPS literature shows inconsistent frequency ranges, heterogeneous block lengths, variable or unreported eye state control, and distances that span 20-30 cm or are not reported. A few recent studies defined orderly sweeps and rests, but intensity remains the most common omission despite clear evidence that these parameters drive photic responsiveness. Until there is convergence on distance, an awake state, and reporting of eye state sequence, frequency list, block and rest timing, and luminance intensity, detection rates will remain study-specific, and standardization will be limited.

Influence of vigilance state on IPS yield

The benefit of sedation and anesthesia is that it improves technical quality by suppressing movement artifacts on EEG.³⁷ However, a consistent trend emerges that both of these medicated states blunt the incremental diagnostic value of IPS for eliciting PPRs in routine EEG in dogs. For example, when propofol was used as a general anesthetic on beagles before testing, IPS neither provoked PPRs nor altered background activity, despite artifact-free signals.³⁷ A follow-up study extended these findings into the diagnostic domain. Under the effect of propofol, EEG in epileptic dogs detected EDs in a quarter to a third of cases, whereas IPS failed to induce new EDs during stimulation.³⁸ Furthermore, medetomidine also was found to provide similar results, and IPS failed to provoke increased EDs in dogs (Table 1).³⁶ Pharmacologically, these 2 agents suppress visual responsiveness through distinct mechanisms. Propofol is a γ-aminobutyric acid type A receptor agonist that potentiates chloride channel conductance, directly suppressing thalamocortical relay neuron excitability and blunting synaptic transmission at the level of the thalamus.⁴⁴ By contrast, α-2 adrenergic agonists, such as medetomidine, act primarily on the locus coeruleus, suppressing noradrenergic arousal pathways and inducing a sedative state resembling natural sleep.⁴⁵ Both pathways converge on decreased thalamocortical transmission, but through fundamentally different receptor-level mechanisms of action. This distinction matters because it suggests that PPR suppression under these agents is not merely a consequence of the sedated behavioral state, but reflects direct pharmacological disruption of the neural circuits required for photic responsiveness.^44,45

These results can further be explained by electroretinograms, which measure the retinal response to light. Propofol, medetomidine, and other common depressants abate visually evoked responses such as a- and b-waves.⁴³ In response to light, a-waves display how strong and fast retinal input is, whereas b-waves display how effectively that input is relayed forward toward the brain.^43,46 Drugs such as medetomidine and propofol increase latency and lower amplitudes of a- and b-waves, resulting in the brain receiving unsteady, less precisely timed visual drives.⁴⁶ This effect has a direct implication on EEG, because impaired a- and b-wave function weakens the chance of generalized spike-waves or polyspike-waves during IPS on EEG.^47,48 These depressants shift the brain into a synchronized, low arousal state, which further decreases thalamocortical responsiveness and explains why IPS did not promote EDs in previous studies.^{36–38,43,46,47}

Overall, it can be concluded that sedation and anesthesia dampen thalamocortical excitability. Although technical feasibility may be convenient through reduction of artifacts and ease of testing, the physiological implication does not translate into high IPS yield for PPRs or improved classification when patients are medicated. Interestingly, there could be pathology-driven exceptions that elicit PPRs under sedation, such as Lafora disease. Under the effects of medetomidine, IPS reliably triggered generalized polyspike–slow-wave complexes with correlated myoclonic jerks after brief latencies in dogs.³⁹ This observation suggests that in strongly photosensitive syndromes, underlying network hyperexcitability can overcome some sedative dampening, allowing IPS to retain diagnostic utility despite sedation. Indeed, the IPS response in Lafora disease is not merely a useful diagnostic adjunct but a consistent and defining electroclinical feature of the syndrome. Characterizing this response represents a step toward formal description of veterinary epilepsy syndromes, a translational goal that the current lack of standardized IPS protocols has so far impeded.

In contrast, awake IPS produces robust electroclinical signals in dogs and preserves native visual and thalamocortical responses. In Rhodesian Ridgebacks with juvenile myoclonic epilepsy syndrome, an awake EEG protocol with added IPS successfully triggered electrographic responses.⁴¹ This resulted in generalized spike-wave and polyspike-wave complexes, all of which coincided with myoclonic jerks.⁴⁰ Interestingly, follow-up work of the same Rhodesian Ridgeback syndrome documented electroclinical features such as absence phenomena, meaning IPS does not always trigger a PPR in every individual.⁴¹ This finding emphasizes the value of awake testing to document both spontaneous and stimulus-linked activity across a photosensitive spectrum. Comparatively, as with Lafora disease, the IPS response in juvenile myoclonic epilepsy of Rhodesian Ridgebacks constitutes a defining electroclinical feature rather than an incidental finding. These breed-specific photosensitive syndromes represent the earliest candidates for formal veterinary epilepsy syndrome classification informed by standardized IPS, directly advancing the translational ambition outlined in the introduction of this review.

Older clinical work also integrated IPS into awake recordings in dogs and demonstrated that photic testing could be used within routine EEG to characterize activity.³⁵ Awake IPS also has shown practical diagnostic value in clinical neurology. A recent first-line awake video-EEG study in dogs and cats reported a PDR in a subset of recordings from dogs and epileptic anomalies during photic testing in 13% of dogs that underwent IPS, confirming that the approach is both feasible and clinically informative.⁴² Awake EEG may have promoted movement and myogenic artifact, but this situation is only a methodological challenge, not a reason to abandon an opportunity where IPS should be implemented. The literature shows that awake dogs can be recorded successfully with appropriate immobilization and video monitoring.⁴² Moreover, modern EEG protocols add electromyography channels to separate myogenic artifact from cortical PDRs.⁴²

Thus, sedation and anesthesia do more than merely quiet muscles, but delay and attenuate visual pathway signals from retina to cortex. These drugs prolong early flash-evoked latencies and delay polysynaptic components, which directly lowers the probability that IPS will elicit generalized spike- or polyspike-waves. Retinal physiology in dogs further supports awake, photopic testing and should be controlled rather than pharmaceutically flattened.⁴⁹ On the whole, the evidence supports an awake IPS protocol as the most faithful test of photic responsiveness, offering the ability to unmask photosensitive physiology when it exists, and avoiding the risk of false negatives by design.

Limitations of spatial sampling in veterinary EEG

Veterinary IPS studies have largely relied on sparse EEG montages, which limits the ability for topographical localization of EDs across the brain. Electrode counts and coverage vary widely, which weakens visualization of the origin of the photic responses and how they propagate through the cerebral cortex (Table 1). For example, a large, anesthetized cohort of dogs was recorded using 5 electrodes placed in frontal, central, and occipital zones of the brain, which omitted parietal and temporal coverage entirely, thus narrowing the view of posterior-to-anterior spread during photic testing.^37,38 The electrodes and outcomes were organized for global visual reads and band power summaries rather than for regional mapping, and any occipital maxima or midline predominance could not be described with confidence.^37,38 No new paroxysmal activity appeared during IPS, but the spatial sampling was too limited to test whether the PDR peaked over visual regions or shifted anteriorly during generalized discharges. Without consistent coverage of occipital and parietal sites, topographic features that define photosensitive physiology remained undetected. Similarly, in a large, sedated outpatient cohort of dogs, 8 subdermal wires were placed over the frontal, central, temporal, and occipital regions only; excluding explicit parietal electrodes led to limited topographic resolution around the occipital cortex, where a PDR is primarily generated.^36,50 The PDR is an occipital phenomenon, and propagation to parietal or frontoparietal networks is atypical and may reflect pathology.⁵¹ In dogs, however, cortical correspondence varies with skull conformation across different breeds. This situation is illustrated by craniocerebral mapping in mesocephalic dogs, where the parietal Pz electrode was found to overlay the occipital lobe.⁵² Because of this anatomical displacement, parietal leads such as Pz can inadvertently record underlying visual cortex activity, emphasizing the need for parietal leads to be incorporated within EEG montages in dogs. Although wire arrays were applied consistently in the large sedated cohort of dogs to support visual scoring and background categorization, the montage did not support regional analyses of IPS effects.³⁶ Results were summarized by ED morphology across pre-photic, photic, and post-photic blocks, which collapses any spatial gradients that might distinguish a PPR range from nonspecific activation.³⁶ Because of suboptimal electrode density and placement for occipital mapping, frequency-specific responses, and topographic maxima or thresholds by region could not be expressed.⁵³ That choice limited both the sensitivity to detect IPS-linked fields and the ability to compare IPS yield across brain centers. Not all veterinary arrays lack parietal coverage. A previously described electrode array does record from parietal regions and is therefore capable of capturing propagation from occipital to parietal cortex.⁹ These results represent a meaningful advance over arrays that omit parietal positions entirely. Furthermore, breed-specific awake state work demonstrated clear electroclinical photosensitivity, but stopped short of spatial mapping. The limited electrode coverage and reporting prevented characterization of regional maxima and propagation patterns during IPS, making it difficult to identify spatial biomarkers that a standardized IPS protocol for dogs should be designed to detect.^40,41 A similar picture appears in photosensitive Lafora disease. Electroclinical responses were robust using a 10-channel 10-20 style montage with bilateral parietal leads, allowing clearer visualization of generalized polyspike-slow-waves across posterior regions.³⁹ However, the montage and analysis emphasized timing and morphology over spatial distribution. The provocative frequency band was identified, but the lack of topographic mapping prevented questions about occipital dominance or interhemispheric symmetry from being answered in a reproducible way. This design limits how these compelling cases can inform a breed-level, clinic-wide, or comparative IPS standard.³⁹ The trend of coarse description of EEG continues to limit comparison across studies regarding spatial resolution and regional analysis. Some studies used a 6-channel bipolar setup with approximately 8 subdermal needles placed, whereas other studies used a 15-needle referential montage that included parietal positions and more.^35,54 These studies are valuable for EEG feasibility, but the sparse and heterogeneous spatial sampling means their yields cannot be used to define topographic criteria for photosensitivity.

Sparse electrode montages limit what veterinary EEG can identify photic responses because they blur where activity begins and spreads. Electrode arrays vary across at least 9 different layouts, and few centers have used systematic topographic coverage.^52,55 This variability weakens lateralization, decreases sensitivity to occipital maxima during PDR, and complicates any attempt to map generalized spike and wave morphology across regions. Standardized EEG placement requires validated arrays, which are a prerequisite for reliable topographic analysis of IPS effects.¹⁰ Modern EEG practice in dogs is beginning to address these shortcomings, because standardized veterinary approaches aligned with recommendations used in humans demonstrate diagnostic feasibility.^{9,42,52,56,57} Electroencephalography in humans methodologically outlines the features of an ideal array for dogs and calls for consistent spatial sampling that supports robust interpretation of activation procedures such as IPS.^52,55 Together, these studies show a path toward reproducible arrays that can support topographic mapping of photic responses in clinical cohorts. Consensus standards specify how IPS should be delivered and reported, and they quantify incremental yield, which depends on the ability to recognize generalized patterns and regional maxima with confidence.^11,25,29 In clinical practice in humans, the standard 10-20 system uses 21 electrodes to provide reliable whole-scalp coverage.²⁵ Veterinary studies, by contrast, have employed as few as 5 and as many as 15 electrodes, with no agreed upon minimum.⁹ Higher electrode counts offer finer spatial resolution, improving ability to localize PDR fields, distinguish regional from generalized discharges, and detect topographic features that sparse arrays may miss. Therefore, arrays at the higher end of those used in current veterinary practice would better support the spatial analysis that IPS interpretation demands. However, caution is warranted in treating electrode counts used in humans as a direct benchmark for EEG in dogs. The spatial extent of the canine cerebral cortex is considerably smaller than that of the human cortex, and EEG is inherently a technique of low spatial resolution. Adequate whole-scalp coverage in dogs therefore may be achieved with fewer electrodes than the 21 used in the 10-20 system used in humans. The recommendation is not to match electrode numbers used in humans, but to ensure that whichever array is used provides consistent coverage of occipital, parietal, temporal, and frontal regions with sufficient density to distinguish regional from generalized discharges. To achieve this rigor, veterinary epileptologists must report defined electrode arrays, document coverage over visual regions, and align IPS outcomes with spatially explicit measures rather than only aggregate counts.

Overall, sparse and inconsistent montages across studies implementing IPS have affected their ability to be incorporated into routine EEG protocols. The net effect is uneven spatial sampling across studies, where some montages have only 5 channels,^37,38 and others have 8, 10, or 12, respectively.^35,39,40 That inconsistency constrains topographic interpretation of IPS responses, risks false negatives for PPRs, and makes standardization across veterinary cohorts difficult. The consequence is that positive findings tend to be expressed as global generalized discharges, and negative findings often are interpreted without testing for regional PDRs over the visual cortex. A modern veterinary protocol should specify a repeatable electrode set that includes robust occipital and parietal coverage, report channel locations, and analyze IPS with frequency-specific thresholds.

Discussion

Our review demonstrates that IPS in dogs is promising, but inconsistent methods and incomplete reporting limit its effectiveness. Awake IPS can identify electroclinical photosensitivity and generalized spike-wave and polyspike-wave activity when it exists, whereas sedation and anesthesia tend to attenuate visual drive and lower the chance of capturing EDs. The literature also has wide variability in frequency ranges, distance, eye state control, and electrode coverage, which makes cross-study comparisons difficult and blunts the ability to set practical thresholds. These findings support the core claim that IPS has clinical value in veterinary EEG, but that its diagnostic yield, interpretability, and comparability are impaired by a lack of standardized stimulus design.

What a standardized veterinary IPS study should implement

Rather than directly adopting an IPS protocol designed for humans, it should be considered a methodological reference point while initially exploring and systematically describing variations specific to dogs (Table 2). This approach would allow comparison across studies and facilitate evidence-based convergence toward an optimized veterinary protocol. Only through structured comparison of different stimulus parameters in dogs can it be determined which elements of the framework designed for humans translate directly and which require species-specific adaptation. Veterinary studies should use the most up-to-date awake video-EEG protocol to ensure the EEG montage samples all parts of the cerebral cortex to support regional outcomes rather than only global counts.^9,57 Synchronized video is crucial throughout stimulation to identify photomyoclonic responses, which can mimic EDs on EEG if not correlated with visible muscle activity.^12,13 The photic stimulator lamp should be placed less than 30 cm from the eyes, with the exact distance documented, and the device’s light intensity setting recorded. Current veterinary IPS protocols are constrained by photic stimulators designed for clinical use in humans, and lamp specifications vary considerably across manufacturers. Recording the lamp type and model is therefore essential for reproducibility, because spectral output and flash characteristics differ among devices. Separate flash trains should be delivered at each frequency for a minimum of 7 s, with a minimum 5 s rest interval between steps.^20,24 Moreover, if a generalized PPR is observed at a given frequency, a rest period of 10 s is recommended before proceeding to the next frequency, allowing cortical excitability to return to baseline and decreasing the risk of cumulative provocation.²⁴ In practice, dogs may move during stimulation, effectively decreasing the duration of uninterrupted photic exposure. A train duration longer than the 7 s minimum is therefore advisable to ensure that sufficient artifact-free stimulation is captured at each frequency. Frequencies should be presented in ascending and then descending order, covering ranges that span the expected sensitivity window between 1 and 60 Hz.^20,24 The precise frequency sequence recommended for IPS in humans (1, 2, 8, 10, 15, 18, 20, 25, 40, 50, and 60 Hz) should serve as the reference standard, not a strict sequence, to enable data modification specifically for dogs.^24,25 In humans, testing 3 different eye states: eye closure, eyes closed, and eyes open, in a consistent order, increases the likelihood of PPR provocation. However, this approach cannot be directly extrapolated to dogs as, realistically, dogs can only be tested with eyes open, with spontaneous blinking expected throughout the procedure. Reporting the full frequency series, exact incremental order, and defining time-locked endpoints, including lower and upper PPR thresholds, should be reported to define a provocation range for each subject. A clear stop rule is essential because IPS can escalate from benign PDRs to generalized discharges in photosensitive syndromes. Technical guidance for humans instructs the operator to stop the stimulus as soon as generalized EDs appear, which requires informed observers during the procedure.^20,24 The timing of IPS within the overall EEG recording session also should be documented, because cortical excitability may fluctuate over the course of a recording. Moreover, a previously proposed grading system as well as common patient IPS characteristics should be adapted to track PPRs and their spread throughout the brain.^12,13,19 Doing so establishes an ethical ceiling for stimulation while still allowing measurement of sensitivity. Together, these steps turn IPS from a single coarse activation into a reproducible assay of photosensitivity that others can audit and replicate.

Table 2.

Recommended reporting checklist for IPS studies in dogs.

Parameter	What to record
EEG montage	Number and position of electrodes; reference type; coverage of occipital, parietal, temporal, frontal regions
Video monitoring	Whether synchronized video used; camera position relative to dog
Lamp	Manufacturer, model, spectral output type (LED, etc)
Lamp distance	Distance from lamp to eyes (cm); how distance was maintained
Light intensity	Luminance or intensity setting used
Frequency range	Lowest and highest frequencies tested (Hz)
Frequency increments	Step size (Hz) and order (ascending, descending, or both)
Train duration	Duration of each flash train (s)
Rest interval	Duration of rest between trains (s)
Eye condition	Eyes state; ambient room lighting
Vigilance	Awake, sedated, or anesthetized; if sedated: drug, dose, route, time
Timing of IPS	When IPS was performed within the overall EEG session
Stop rule	Whether a stop rule was applied
Photomyoclonic response	Whether photomyoclonic responses observed; video correlation
PDR	Whether PDR observed and at what frequency
PPR	Whether PPR observed and topography

Abbreviations: EEG = electroencephalogram; IPS = intermittent photic stimulation; LED = light-emitting diode; PDR = photic driving response; PPR = photoparoxysmal response.

Remaining unknowns in canine photosensitive epilepsy in dogs

In people, specific spectral color and flicker properties are well described, for example, red-rich stimuli and high-contrast patterns are strongly provocative of PPRs.^31,32 Moreover, red and blue colored lens filters can abolish or decrease PPRs in many patients.^31,32 In contrast, virtually no work has systematically mapped seizure susceptibility across different light spectra for dogs. Color perception in dogs is dichromatic with cone sensitivities clustered roughly in the violet-blue and yellow ranges.^30,33,34 This suggests that both the color space and light frequency of provocative stimuli will differ from data obtained in humans. Future IPS studies in dogs should compare cortical responses to narrow band light targeting wavelengths to which dogs are sensitive, such as shortwave blue light near 430 nm and yellow-green light near 555 nm against standard broadband white light.⁵⁸ Because narrowband stimuli matched to canine cone peaks would concentrate energy where the retina is most responsive, it may prove more effective at evoking PDRs or PPRs. Applying a similar wavelength-mapping approach seen in humans to dogs, which could help further characterize photosensitive epileptic syndromes. Therefore, extending this approach from retinal to cortical endpoints via EEG would represent a logical next step and facilitate the development of a comprehensive, species-specific IPS protocol.

Similar uncertainty surrounds video and screen-based triggers. Reviews in humans distinguish an array of video screens but emphasize that it is usually the content, such as intense color contrast, that provokes seizures rather than the device itself.¹⁷ In dogs, dedicated studies linking televisions, monitors, LED toys, or video content to PPRs is essentially absent. Current EEG and IPS literature in dogs does not characterize responses to realistic screen-like stimuli, and there are no controlled data on safe refresh rates, brightness, or viewing distance for at-risk breeds.¹⁰ It is possible that PPRs would be triggered by entertainment media or environmental triggers such as flash photography but, because dogs generally are positioned at viewing distances that decrease potential risk, this assumption remains speculative. That said, the observation that PPRs occurred in Rhodesian Ridgebacks exposed to a television screen suggests that ambient photic stimulation at a distance can provoke cortical responses in photosensitive dogs.⁴⁰ This raises a practical question for noncompliant or aggressive dogs. Rather than resorting to sedation, researchers potentially could deliver flashes in a darkened room, with the dog positioned at a safe distance from the operator. Such an approach would preserve the awake vigilance state while decreasing handling-related stress and risk.

Beyond flash-based stimulation, patterned visual stimuli also may warrant investigation. In humans, striped patterns with spatial frequencies near 3 cycles per degree provoke maximal visual cortical activation and can induce seizures in photosensitive epileptic individuals.⁵⁹ Whether similar spatial frequencies would provoke cortical hyperexcitability in dogs is unknown because no equivalent investigation has been conducted in this species. Future studies could explore pattern-based stimulation as a complement to conventional flash-based IPS, because it may identify an additional dimension of visual sensitivity in dogs with epilepsy.

Overall, future work in photosensitivity of dogs should first prioritize rigorous standardization of IPS protocols. Only after a consistent, reproducible framework for eliciting and identifying PPRs has been established will it be meaningful to systematically interrogate secondary variables such as spectral color, screen-specific features, and other environmental modifiers in photosensitive dogs.

Conclusion

Standardized awake IPS should become the most reliable way to test photic responsiveness in dogs undergoing EEG. It preserves native visual and thalamocortical physiology, allows time-locked measurement of PPR thresholds, and supports spatial interpretation when montages include robust cortex coverage. The field now needs a harmonized stimulus design that enables IPS to evolve from selective use to a reliable activation in veterinary practice.

Abbreviations

EEG

electroencephalography

ED

epileptiform discharge

IPS

intermittent photic stimulation

LGN

lateral geniculate nucleus

PPR

photoparoxysmal response

Author contributions

Robert John Shwarzman (Conceptualization, Writing—original draft, Writing—review & editing), Marnie Ford (Writing—review & editing), Alex Chan (Writing—review & editing), Lluis Gaitero (Writing—review & editing), and Fiona M. K. James (Conceptualization, Funding acquisition, Supervision, Writing—review & editing)

Conflicts of interest

The authors declare no conflicts of interest.

Funding

We acknowledge the support of the Natural Sciences and Engineering Research Council of Canada (NSERC), funding reference number RGPIN-02606-2021.

Data availability

No new data were generated or analyzed in this study.

Off-label antimicrobial declaration

The authors declare no off-label use of antimicrobials.

Institutional animal care and use committee or other approval declaration

The authors declare no institutional animal care and use committee or other approval was needed.

Human ethics approval declaration

The authors declare human ethics approval was not needed.