Neurostimulation in Neuro-Ophthalmology: Mechanisms and Therapeutic Potential

Abstract

Visual processing constitutes a substantial portion of cognitive, executive, and sensorimotor functions of the brain. Understandably, damage to visual areas and pathways results in various impairments. Neuro-ophthalmology addresses these complexities, yet traditional management approaches often have limited efficacy and undesirable side effects. In recent years, neurostimulation has emerged as a promising alternative, offering strong therapeutic benefits with minimal adverse effects. While extensively explored in neurological and psychiatric disorders, its application in ophthalmology remains relatively underexplored. This chapter navigates recent advancements in neurostimulation techniques, focusing on their potential in treating neuro-ophthalmic illnesses. We begin with an introduction to the visual system and then cover major neuro-ophthalmologic illnesses and related stimulation principles while also describing associated neurochemical and neuroplastic changes. Two major types of neurostimulation modalities in ophthalmology are discussed—invasive and non-invasive—highlighting their mechanisms and therapeutic potentials. Finally, we address current challenges, gaps, and prospects in neurostimulation research in ophthalmology in managing neuro-ophthalmic disorders.

1. Introduction

We, like many animals, are primarily visual creatures, and processing visual information for cognition and behavior occupies a sizable section of the cerebral cortex [1]. About 30% of the cerebral cortex is dedicated to visual processing in conjunction with numerous additional non-visual areas dispersed all over the brain [2, 3, 4]. Therefore, the loss/damage of visual areas and pathways may lead to several perceptive, cognitive, spatial, memory-related, sensorimotor, and other declines in behavior [5, 6].

Neuro-ophthalmology is a special branch of medicine that deals with a variety of complex visual issues caused by neurological conditions, thus posing significant challenges to the affected individual [7]. Despite several developments in traditional management strategies (pharmacological or surgical) for various neurological disorders to control such conditions, the undesirable adverse effects and low treatment response rates have paved the way for alternative modern techniques [8, 9, 10, 11]. Neurostimulation, the process of modifying or modulating nerve activity by administering electrical stimulants directly to a specific brain region, is a rapidly growing technique with few adverse effects and strong therapeutic benefits [12]. Notably, these techniques are adaptable and customizable contingent upon the subject [13] and can be applied as a single treatment, by constantly stimulating the region (on a duty cycle with specific parameters), or in response to physiological changes [14], with applications extending to several diseases/impairments [15]. Additionally, it can also be directed at a malfunctioning brain area or network in neurological and psychiatric disorders, such as dystonia [16], obsessive-compulsive disorder [17], focal epilepsy [18], bipolar disorder [19], schizophrenia [20], chronic pain [21, 22], and depression [23].

Neurostimulation in ophthalmology is rather underexplored for its therapeutic potential. In this chapter, centered around the advances, mostly over the past few years, we navigate it as a promising prospect in neuro-ophthalmic illnesses. We first introduce different types of stimulation and the principles of stimulation (with focus on ophthalmology). We then overview the general visual perception pathway, followed by a description of some major neuro-ophthalmologic illnesses. We then discuss neurochemical changes and mechanisms along with neuroplastic changes, followed by two major types (invasive and non-invasive) of neurostimulation modulations in ophthalmology. Finally, we discuss challenges and gaps with future forecasts in ophthalmological research and applications.

2. Types of neurostimulation: invasive and non-invasive

Neurostimulation is classified into two types: invasive and non-invasive techniques. While the former requires the insertion of an electrode or stimulation tool inside the specific area of interest, the latter is carried out externally, making it safer and easier for applicability [24, 25], but with inadequate spatial resolution and shallow penetration [26]. The invasive techniques primarily include visual implants/prostheses and deep brain stimulation (DBS). On the other hand, the non-invasive techniques fall into several categories, such as transcorneal electrical stimulation (TES), transcranial magnetic stimulation (TMS), transcranial direct current stimulation (tDCS), transcranial alternating current stimulation (tACS), random noise stimulation (RNS), transcranial ultrasound stimulation (TUS), and vagus nerve stimulation (VNS) [27, 28, 29].

In neuro-ophthalmology, depending on the stimulation site, stimulation falls into two categories: pre-chiasmatic and post-chiasmatic [30]. The pre-chiasmatic stimulation includes corneal (transcorneal alternating current stimulation, tcES), eyelid (transpalpebral alternating current stimulation, tpES), and peri-orbital zone (repetitive transorbital alternating current stimulation, rtACS) stimulations. These are performed to treat optic neuropathy, retinitis pigmentosa, glaucoma, retinal artery occlusion, macular degeneration, and others [31]. The post-chiasmatic stimulation comprises tDCS and high-frequency random noise stimulation (hf-RNS) to treat brain disorders, such as hemianopsia and brain damage. Notably, these are also delivered to treat pre-chiasmatic diseases, such as amblyopia and myopia [30, 32, 33].

3. The visual pathway: overview and organization

The visual cortex is one of the most studied areas of the brain, and we have a relatively good understanding of the architecture, morphology, physiology, and functional organization of the visual pathways across many species [34]—a fundamental premise to comprehend the outcomes of electrical stimulation of the visual cortex in relation to different eye diseases. The visual pathway (and perception) begins with the globes (eyeballs), extending all the way to the cortex.

3.1 The optic pathway

The optic pathway begins in the retina, a complex structure composed of ten distinct layers, each with its own function. The photoreceptor layers, which include rods and cones, generate action potentials through photosensitive cycles involving rhodopsin.

The ganglion cell layer and the nerve fiber layer form the basis of the optic nerve; the ganglion cell layer contains cell bodies, while the nerve fiber layer contains axons that extend across the retina [35, 36]. These axons consist of two types of fibers: temporal fibers, which control the nasal visual field, and nasal fibers, which control the temporal visual field. These fibers converge at the optic disc and exit the eye posteriorly to form the orbital portion of the optic nerve. The nerve is encased in the dura mater, continuous with that of the brain, allowing cerebrospinal fluid to move freely between the eye and the intracranial space. The axons leave the orbit through the orbital foramen along with the ophthalmic artery and sympathetic fibers [35, 36]. The retina of the eye has two types of photoreceptor cells: rods and cones that are specialized neurons with photosensitive proteins that are uniquely activated by the different wavelengths of light [37, 38]. The fovea is the part of the retina where the acuity of the image is the sharpest and only contains cones (responsible for vision in high-intensity light during the day), whereas the peripheral retina has both rods (responsible for vision in low-intensity light during the night) and cones [39]. The information is then relayed to the retinal ganglion cells (RGCs) through bipolar cells and finally reaches the optic nerve (the second cranial nerve; CN II) [36].

Figure 1A displays the parts of an eye. As the light from an object hits the cornea, it then is focused by the pupil and the lens to relay visual data from the surrounding environment onto the innermost nervous layer of the eyes, the retina — the eye’s neurosensory structure, where the image of the object is first represented in an inverted fashion [40]. Through the optic canal, the optic nerve exits the orbit and travels to the optic chiasm, where it converges with the contralateral nerve in the suprasellar cistern [41]. Figure 1B illustrates the pathway from the eye to the retina and to the cortex. Some medial axons from each optic nerve decussate (cross over) joining the lateral axons at the contralateral side [42]. In this way, the right and the left visual fields are represented in the brain; that is, the information coming from the left side of the visual scene mostly enters the right part of the brain and vice versa [43].

Figure 1.

From the eyes to the visual cortex. (A) Illustration of the anatomical structure of the eye, spanning from the cornea to the optic nerve. The purple triangle represents an object (external light source) that is first represented as an inverted image on the retina (fovea, the region of highest acuity on the retina). (B) Shows the right (red) and left (blue) visual hemifields. The decussation of the optic nerves at the optic chiasm results in the processing of the right visual field by the left hemisphere. Similarly, the left visual field is processed and represented by the right hemisphere. The visual cortex is organized into retinotopic maps, i.e., sensory space is systematically organized in the cortex. Note: This figure is an original figure created by the authors.

3.2 Retinogeniculate and retinotectal pathways

From the retina, the information is relayed to the lateral geniculate nucleus (LGN) in the thalamus (retinogeniculate pathway) or directly to the superior colliculus in the midbrain (retinotectal pathway) [44]. The optic radiations, composed of second-order neural cells, originates from the thalamus and relays the information to the visual cortex [45]. Thus, the visual information reaches the extrastriate cortex via these two separate pathways [46].

The retinogeniculate pathway (RGVP) travels through the optic nerve, crosses at the optic chiasm, continues along the optic tract, and ultimately reaches the LGN [41, 47]. Around 53–58% of RGVP fibers cross over at the optic chiasm [48, 49, 50]. This results in the RGVP being categorized into four anatomical segments: two that cross (medial) and two that do not cross (lateral), with each segment responsible for transmitting visual information from visual hemi-fields [48, 51, 52]. Studying the RGVP is crucial for various research and clinical purposes. Many diseases can impact the RGVP, such as pituitary tumors [53], glioma [54], optic neuritis [55, 56, 57], ischemic optic neuropathy [58, 59], and optic nerve sheath meningioma. Specifically, visualizing the RGVP is beneficial for planning surgical approaches to both intrinsic and extrinsic lesions of the pathway [60].

On the other hand, the retinotectal pathway through the superior colliculus processes this information to coordinate eye movements and direct attention to visual stimuli, thus playing a crucial role in visual orientation and reflexive responses [61, 62]. Neurostimulation of this pathway can potentially enhance or restore visual function, particularly in individuals with visual impairments. Techniques such as transcranial magnetic stimulation (TMS) or optogenetics can modulate the activity of neurons in the superior colliculus, influencing visual processing and eye movement control [24, 63, 64, 65, 66].

3.3 Visual streams: dorsal vs. ventral

Different regions of the visual cortex in the brain have specialized functions related to processing visual information [67]. The primary visual cortex (V1) in the occipital lobe is the first part of the cortex where the information arrives for primary interpretation [68]. V1 plays a crucial role in identifying spatial frequencies, orientation, and color [69]. V2, another area of the visual cortex, supplements V1’s functions by assisting in the perception of contours [70]. Area V1 receives input primarily from the lateral geniculate nucleus (LGN) and contains ocular dominance columns (ODCs) which facilitate spatial frequency recognition and contribute to binocular vision [71, 72]. The V1 comprises functional modules where neurons with similar receptive fields align into columns [73]. From V1 onwards the visual information travels parallelly in two separate [68] yet interconnected [73] pathways: the dorsal and ventral pathways (Figure 2).

Figure 2.

Dorsal and visual pathways. The information from the retina to the V1 takes two separate pathways: through LGN (retinogeniculate, orange) and through SC (retinotectal, yellow). From V1, the visual information flow is divided into the dorsal (where pathway, blue) and ventral (what pathway, red) pathways, with cross-interactions between them (magenta arrow). These pathways also interact with the frontal cortex (green arrow). This network suggests that loss/damage to any area(s) along these pathways may cause loss of multiplicity of functions: visual, cognitive, motor, and others [68, 73]. Note: This figure is an original figure created by the authors.

The ventral pathway (what) begins with V1 relaying information in a hierarchical manner to the secondary (V2) and other associative areas (V3, V4), terminating in the inferior temporal (IT) cortex [74, 75, 76]. Notably, the receptive fields (area of space that modulates neural activity) project specifically from one area to another forming contralateral retinotopic maps, thereby integrating simple information into a complex percept [74]. Conversely, the dorsal pathway (where pathway) branches into the dorsal part of the brain from V3 onwards to the medial temporal (MT, also called V5), terminating in the superior parietal lobule (SPL) [77, 78]. Collectively, as the names suggest, the ventral (what pathway) is responsible for detecting and combining different characteristics (orientation, color, spatial frequency, temporal frequency, direction, etc.) [79, 80, 81], whereas the dorsal pathway (where) encodes the spatial location and action [82, 83, 84].

4. Disorders of the eye

Eye disorders are classified according to the part of the eye affected. Below, we cover some major orders across both types.

4.1 Retinal disorders

Photoreceptor degeneration, which includes both structural and functional deterioration of the cones and the rods [85], leads to a remarkable reduction in the ability of the retina to perceive light, ultimately resulting in subsequent vision loss. This may end up with retinal remodeling, an event cascade that permanently changes the neural retinal structure due to an altered retinal environment [86].

Retinal remodeling is often described in three phases in which multiple cellular and molecular events occur in an attempt for injury adaptation [85, 87, 88]. The early phase of retinal remodeling begins with photoreceptor stress. The stress on these cells leads to the dislocation of photoreceptor opsins, which are light-sensitive pigments found within rods and cones (types of photoreceptor cells). This dislocation activates cellular reactions that cause the photoreceptor cells to hyperpolarize (change their electrical charge), and their outer segments move into inner adjacent retinal layers, disrupting the normal organization of the retina [85]. This disruption results in impaired transmission of visual signals, affecting vision. Photoreceptors’ death with subsequent activation of microglial cells [89] results in an even greater loss of photoreceptors [90] and bipolar deafferentation. Müller cell hypertrophy [89, 91], an enlargement or swelling of Müller cells, a type of glial cell found in the retina leads to the separation of the neuronal retinal layer from the retinal pigmented epithelium (RPE) and choroid [92]. The RPE and choroid are layers of tissue beneath the retina that provide essential support functions, including blood supply, nourishment, and immune regulation, which are crucial for the health and function of the retina. Briefly, Müller cell hypertrophy disrupts the normal relationship between the neuronal retinal layer and the supporting layers beneath it, potentially impacting the retina’s blood supply, nourishment, and immune regulation, which are essential for its proper functioning [92]. In the chronic phase, persistent remodeling with gross rewiring, de-novo axonal formation, sprouts from all the remaining neurons [93]. Combined with neuronal migration, it leads to restructuring of the retina, irreversibly altering retinal architecture and function, thus causing long-term progression of visual loss [93].

Two conditions mostly contribute to the photoreceptor degeneration: age-related macular degeneration (AMD) and retinitis pigmentosa (RP). AMD is typically acquired later in life and is characterized by the accumulation of yellowish deposits called drusen under the retina [94]. These deposits accumulate in the atrophic (degenerated) areas of the RPE, especially at the macula, the central region of the retina responsible for detailed central vision [94]. The presence of drusen and the atrophy of the RPE can lead to a progressive loss of central vision over time, as the function of the photoreceptors becomes compromised. AMD is a leading cause of vision loss in older adults and can significantly impact the quality of life. There are two main forms of AMD: dry (non-neovascular) AMD, characterized by the presence of drusen and RPE atrophy, and wet (neovascular) AMD, characterized by the growth of abnormal blood vessels beneath the retina, which can lead to sudden and severe vision loss if left untreated [95]. RP is a group of rare genetic disorders that affect the retina. These disorders typically have an early onset and may initially affect the peripheral vision before progressing toward the center of the retina [85, 96, 97]. It is associated with defects in various components of the retina, including opsins (the light-sensitive proteins found in photoreceptor cells), the RPE, or other molecules involved in retinal function [98]. Despite the specific molecular defects involved, the common outcome in RP is the degeneration of photoreceptor cells. As photoreceptor cells degenerate, individuals with RP often experience progressive vision loss, starting with night blindness and peripheral vision loss and eventually leading to central vision impairment or blindness in severe cases. The specific genetic mutations and the rate of disease progression can vary among individuals affected by RP [99].

Diabetic retinopathy is another common retinal disorder and a leading cause of blindness in diabetics [100]. Elevated sugar levels detrimentally affect the neurovascular system in the eye. This condition is categorized based on its severity. In less severe cases, known as non-proliferative diabetic retinopathy, there are observable microvascular indicators such as microaneurysms, macular edema, and hemorrhages [101]. In more severe cases, termed proliferative diabetic retinopathy, there is the growth of abnormal and fragile blood vessels, leading to significant vision impairment due to bleeding in the vitreous cavity and detachment of the vitreous [101, 102].

A class of neurodegenerative illnesses known as glaucoma is marked by damage to the optic nerve and the gradual degeneration of RGCs. The pathophysiology of all types of glaucoma involves the cupping of the optic nerve head and the death of RGCs and their axons [103]. The elevation of intraocular pressure is thought to be the primary cause of the increased apoptosis of RGCs in glaucoma, as it is caused by either increased production or decreased outflow of aqueous fluid [104].

4.2 Optic nerve disorders

Optic neuropathy is a condition caused by any damage to the optic nerve that can lead to a visual impairment(s). It encompasses a wide range of disorders with different reasons (e.g., compressive, ischemic, inflammatory, diabetic, and toxic) [105]. Ischemic optic neuropathy (ION) is a condition caused by a diminished blood supply to the optic nerve resulting in nerve injury. The ischemic insult can be anterior (at the optic disk, AION) [106] or posterior (to the optic disk, PION) [107]. ION can be further classified according to the cause into arteritic (inflammation of blood vessels like giant cell arteritis) [108] or non-arteritic (other causes) [109]. Optic neuritis (ON) is an inflammatory demyelination of the optic nerve. It is associated with autoimmune reactions resulting from either an underlying autoimmune condition (e.g., multiple sclerosis) or other factors [110, 111, 112]. The inflammation may extend to the optic chiasma (chiasmal optic neuritis) that can cause bitemporal hemianopia and impaired vision of the outer halves of both eyes [113].

5. The principle of neurostimulation in ophthalmology

The earliest thorough and extensive studies on electrical stimulation (ES) of the human visual cortex were pioneered by neurosurgeon Wilder Penfield [114, 115]. After individuals had direct cortical stimulation (DCS), Penfield recorded their visual perceptions or phosphenes, and he demonstrated that areas close to the occipital pole were more likely to generate visual perceptions [115].

The majority of neuromodulatory stimulation techniques rely on electricity, but ultrasound and magnetic field stimulation-based techniques also exist [116, 117]. ES is used in various neurological disorders to modulate the neuronal activity and restore the normal physiology of the dysfunctional neural networks for therapeutic purposes [118]. In ophthalmology, three main types of ES are done depending on the mechanism of action. The first category involves the replacement of a damaged area by mimicking its function allowing for downstream stimulation of the intact visual pathway (i.e., visual prostheses such as retinal implants) [119]. The second category involves the alternation of network activity restoring the normal circuit needed for proper visual function (e.g., tDCS, tACS, DBS, and rTMS) [24]. The third category involves activating neural survival pathways in a damaged part of the neurovisual system to improve visual outcomes [e.g., transcorneal alternating current stimulation (TcACS)] [120].

Numerous clinical trials have been conducted investigating the different neurostimulation devices and protocols to optimize the current intensity, electrode positioning, current frequency, and duration. This, in turn, improves the electrode design and surgical techniques for better outcomes; however, there is still no consensus regarding the standardization and optimization of ES protocols [121]. ES can be delivered to many sites along the visuo-cortical pathway. However, the retinal ES implant technique has attracted the attention of many researchers because of lower surgical complications and better accessibility [122]. There are two forms of electrical neurostimulation in use: prosthetics and non-invasive electric stimulation. While using prosthetics, the downstream visual pathways are stimulated by bypassing the less conductive dysfunctional areas [123]. In non-invasive, low-level electrical stimulation (100–1000 μA), one or more electrodes are positioned close to the eye [124, 125]. The invasive approach, however, being precise in the delivery of ES also carries the risks of infection and surgical complications [126]. The non-invasive approach has a safer profile with only mild adverse effects (e.g., itching and tingling) [127], but it has less spatial resolution [128, 129]. Thus, a combination of devices and approaches allows a trade-off giving the patient and the doctor a range of treatment options.

6. Electrical stimulation in degenerative ophthalmic disorders: neurochemical, immunomodulatory, neuroplastic, and other changes in

Multiple basic science studies have explored the cellular and molecular mechanisms of ES (especially the minimally invasive approaches), explaining the improvement of visual outcomes in a variety of eye diseases [31]. Ample evidence from molecular and electrophysiological studies advocates for this technique for several visual disorders. We highlight specific neurostimulation mechanisms in some inherited and acquired eyes diseases:

6.1 Neurochemistry and immunomodulation: neuroprotective roles in different retinal diseases

The widely accepted explanation for the neuroprotective benefits is that ES may both increase the intrinsic sensitivity of neurons to endogenous neurotrophic factors and up-regulate their expression levels. For example, ES-induced Müller cells’ mRNA levels and proteins for vascular endothelial growth factor (VEGF), brain-derived neurotrophic factor (BDNF), ciliary neurotrophic factor (CNTF), and insulin-like growth factor-1 (IGF-1) increase noticeably [130, 131, 132, 133, 134, 135]. Hanif et al. [136] found an upregulation of gene expression of BDNF, caspase 3, (FGF2), and glutamine synthetase (GS) after a 30-min low-level (4 μA) electrical stimulation to a rodent model with RP. These are important factors for triggering the intrinsic survival system [137, 138, 139, 140, 141], thus preventing the systems against damage [142, 143]. Specifically, Kanamoto et al. [144] identified an upregulation of 25 proteins mediating retinal function through signaling, immunomodulatory effects, and axonal and dendritic regenerative processes [135].

Similarly other apoptotic genes are downregulated [143] when applied direct ES to Müller cell culture [132]. Conclusively, several studies point to the fact that BDNF plays a central role in neuroprotection and is upregulated during ES [133, 143, 145]. Moreover, authors [132, 146, 147] have also shown that ES on axotomized RGCs of the optic nerve (invasive/non-invasive) led to an enhanced survival through activation of neuroprotection mechanisms including IGF-1. ES has also shown a rise in GTPase levels modulating calcium signaling and enhancing directed RGC axon growth toward the cathode electrode in a rodent model [148, 149].

Furthermore, it has been shown that ES plays a role in limiting the inflammatory response in some eye diseases, thus playing a significant role in immunomodulation [150]. It interferes with the activation of inflammatory cells and reduces the production of pro-inflammatory cytokines [151, 152]. Microglial cells, the most abundant CNS immune cells, responsible for the innate immunity of the CNS and injury repair are involved in inflammatory response modulation [153, 154]. Notably, depending on the condition, microglial activation may be beneficial or damaging in a variety of brain and retinal disorders [155]. Retinal microglial cells get activated after ocular insults mediating local inflammatory response and triggering pro-inflammatory signaling. Tumor necrosis factor-alpha (TNF-α) is the most prominent cytokine expressed by the activated microglial cells that is responsible for harmful neuroinflammatory and neurodegenerative processes by acting on its receptor (TNFR1) [156, 157, 158]. Additionally, it mediates the activation of nuclear factor kappa B (NF-κB) and c-Jun N-terminal kinase (JNK) pathway, further reactivating the glial cells and production of neurotoxic pro-inflammatory cytokines, eventually causing the loss of RGCs [159, 160, 161, 162]. After TES treatment in rodent models, a significantly lower number of activated microglial cells were reported [120, 145]. This was associated with a decline in pro-inflammatory cytokines expression (IL-6, TNF-α, and COX-2) along with suppression of the NF-κB pathway and upregulation of anti-inflammatory cytokines, IL-10. Similarly, the application of transpalpebral electrical stimulation (TpES) has shown a significant decrease in activated microglial cells with prolonged photoreceptor survival in inherited photoreceptor degeneration mice [163].

6.2 Promotion of neuroplasticity

A fundamental property of neurons is their capacity to reorganize the synaptic network and the strength and efficacy of synaptic transmission through a diverse number of activity-dependent (e.g., when learning a new skill, practice, or forming memory) [164, 165, 166] or training-induced (adaptation, deprivation, and environmental enrichment) [165, 167] mechanisms. These mechanisms are characterized by synaptic redistributions and synaptic scaling, typically referred to as synaptic plasticity [120, 158, 168]. Interestingly, recent research has shown evidence of improvement in visual function after damage to the visual system in response to electrical stimulation, either structurally through neuronal regeneration or functionally via modulating network connectivity and electrophysiological parameters [169, 170, 171, 172, 173]. ES techniques, such as retinal prostheses or TcACS, aim to bypass damaged photoreceptors and directly stimulate surviving retinal cells or higher visual processing centers. By delivering electrical impulses to the retina or the visual cortex, these approaches can induce neuroplastic changes, including synaptic remodeling and functional reorganization, to enhance visual perception and restore vision [174, 175]. Additionally, ES may facilitate the integration of artificial visual inputs with residual visual pathways, promoting adaptation and improving visual outcomes in individuals with various eye diseases [176]. In summary, by enhancing synaptic efficacy and facilitating the formation of new connections, ES can aid in restoring function after neurological damage or promoting skill acquisition and cognitive enhancement in healthy individuals.

6.3 Enhancement of retinal blood flow

The retina is one of the most oxygen-requiring tissues in the human body and maintaining efficient oxygen circulation to the retina is imperative for optimal visual function [177, 178, 179]. Studies have shown that ES can induce blood flow into the retina. Evidence comes from different kinds of electric treatment, such as transcorneal alternating current stimulation (tcACS) and TES, revealing their positive effect on chorioretinal blood flow in healthy individuals [180] and people with eye conditions such as RP [181, 182]. This increased blood flow comes (without any effects on the systemic blood flow) through adjustments in the vascular tone, thus influencing the neurovascular system, but the exact mechanism is not yet well-known [183].

Some studies suggest an upregulation of some growth factors, such as vascular endothelial growth factors (VEGF), which are responsible for the increased chorioretinal blood flow [184, 185]. Other studies point toward ES-induced upregulation of arachidonic acid metabolites and neurotrophic factors released by Müller cells [132, 149, 186, 187]. Additionally, a study has reported the role of nitric oxide (induced by IGF-1 released by activated Müller cells) in mediating vasodilatory effects [188]. In summary, ES leads to increases in blood flow to the eye without the need for surgeries.

7. Electrical stimulation in neuro-ophthalmology: invasive and non-invasive

Several neurostimulatory approaches have evolved over the last few decades. Initially, invasive techniques were developed during the second half of the last century as promising devices for visual field defects [115, 119]. However, in the early 2000s, non-invasive ES offered an easier method with fewer complications when applied with low current tDCS to the visual cortex, thus inducing phosphenes [189].

7.1 Invasive visual stimulation (visual prostheses/bionic eye)

Invasive stimulation approaches are silicon-based bioelectronic devices that help restore sight only in conditions with limited damage along a specific point of the visual pathway while the rest of the visual system is intact. Figure 3 highlights these visual areas. These are composed of an electrode array (a group of electrodes in a grid) producing a sequence of tiny electrical pulses from images taken using an external tiny video camera worn like glasses. Anywhere along the visual pathway, this device can be implanted to provide neural ES, including the retina, the optic nerve, the visual cortex, and the thalamus. Visual restoration using prosthetics is often referred to as “artificial vision”.

Figure 3.

Visual pathway and different visual areas that can be electrically stimulated for prosthetic applications, thus restoring vision. Note: This is a tractography image generated by us; however, the display of the retina (brown) is an artistic rendering of the eyeball.

The underlying principle of artificial vision devices mainly stems from neurophysiologic principles of the visual pathway. One principle is that the electric stimulus can serve as a substitute for light leading to visual perception [115]. Another principle underlines that prosthetic replacement of the impaired hub/point in the visual pathway fills the functional/structural hole left by impaired areas, thus subserving the remaining intact pathway [190]. Additionally, the retinotopic organization of the visual cortex allows for better localization of points for electrical stimulations to produce rational visual percepts [190].

7.1.1 Retinal prostheses

Retinal prostheses are designed for the damaged photoreceptors of retina and mimic the activity of photoreceptors by converting light into electrical impulses, thus activating ablated/silent downstream neural pathways in degenerative diseases such as RP or AMD [191]. Patients can then see their environment (almost like a grid) when phosphenes of different strengths are saliently triggered in different areas of their retinotopic map [192]. In retinal prostheses, triggering (direct or indirect) the ganglion cells is the key objective of ES [193, 194]. In direct stimulation, the ganglion cells are directly activated without any presynaptic impulses from other retinal cells, such as bipolar cells, that send synaptic projections to the RGCs [195, 196]. Somatic stimulation can be achieved by triggering the cell bodies [197, 198], whereas axonal stimulation activates the axons of the ganglion cells that are right under the electrode [199]. On the other hand, the indirect stimulation method entails the activation of other cells in the retina, namely the bipolar cells, amacrine cells, or horizontal cells [200, 201]. These cells then excite the ganglion cells, thus ameliorating the visual condition by bridging the broken pathway. Table 1 summarizes the main eye diseases and their ES techniques.

Technique	Functionality	Implantation Sites	Components	Benefits	Limitations
I. Invasive
Retinal prostheses for RP, AMD [202, 203, 204, 205]	Mimic photoreceptors’ activity by converting light into electrical impulses, activating downstream neural pathways, and replacing damaged photoreceptors in degenerative diseases, such as RP or AMD.	Epiretinal, subretinal, suprachoroidal locations	Electrode array	Restoration of visual perception in degenerative diseases. Higher resolution and specificity with direct stimulation methods.	Delicacy of retinal structure. Ongoing degeneration in visual pathway. Engineering challenges for high-resolution vision.
Optic nerve prosthesis for RP [206, 207]	Stimulating optic nerve to bypass damaged retina.; Penetrating electrodes provide more focused stimulation.	Extracranial part of optic nerve	Spiral cuff electrodes, penetrating electrodes	Perception of phosphenes across broad visual field.	Less likelihood to impair retinal synaptic processes.; Surgical challenges.
Lateral geniculate nucleus prosthesis for optic neuropathies, retinal degenerative diseases, and eye trauma [208, 209]	Stimulate LGN to restore vision, exploiting its spatial distribution of receptive fields and structural organization.	Lateral geniculate nucleus	Microelectrode array	Less-density electrode array with wider receptive fields. Potential for selective layer stimulation.	Complex surgical procedure.; Proper synchronization between LGNs required.
Visual cortex prostheses for various visual impairments [210, 211, 212]	Stimulate primary visual cortex with penetrating electrode arrays for high-resolution artificial vision.	Primary visual cortex	Penetrating electrode arrays	Higher resolution due to the large surface area of visual cortex. Alternative option for other ineffective prostheses.	Challenges in biocompatibility. Ongoing trials for enhanced encoding. Tolerance by host tissue.
II. Noninvasive
TES for optic and retinal disorders [213, 214, 215, 216, 217]	Electrical stimulation via silver thread DTL electrodes or contact lens on the cornea. Reference electrode placed under the skin.	Cornea	Silver-thread DTL electrodes or contact lens, reference electrode	Improved visual function outcomes in clinical trials. Neuroprotective and therapeutic potential on retinal disorders.	DTL electrodes may cause corneal complications and require careful placement.
rtACS for optic nerve injuries [169, 218, 219, 220, 221]	Delivery of alternating currents through electrodes near supraorbital and infraorbital ridges. Entrainment of brain oscillations to external force.	Supraorbital and infraorbital ridges	Multichannel device delivering sinusoidal pulses through four separate periorbital electrodes.	Enhanced visual outcomes observed in optic nerve injuries. Modulation of oscillatory brain activity, synchronization, and synaptic plasticity.	May cause less corneal complications compared with TcES.
TpES for AMD (TpES) [222, 223, 224].	Application of electrodes on the eyelid, minimizing direct contact with the cornea.	Eyelid	Electrodes on the eyelid	Effective in treating AMD. Minimizes the risk of local eye complications. Helps lower IOP in primary open-angle glaucoma (POAG)	Limited to treating certain eye conditions.
LIFUS for various eye disorders; [225, 226, 227].	Noninvasive stimulation of neurons using low-intensity focused ultrasound.	Retina, visual cortex	Low-intensity focused ultrasound	Higher spatial resolution and deeper penetration. Stimulation of RGCs or visual cortex with potential for vision restoration.	Not yet approved for clinical use. Limited studies on safety and efficacy in humans. Requires development and validation.
tRNS on the visual cortex; [228, 229, 230]	A biphasic sinusoidal alternating current with weak intensity (in mA) in unpredictable random frequencies is applied transcranially targeting the visual cortex	Primary visual cortex	Two electrodes, the active is placed over the visual cortex	Promotes visual perceptual learning in patients with central visual loss	Mechanism of action is not known, thus limiting the ability to improve the protocol and design

Table 1.

Summary of the techniques, principles, mechanisms, targeted disorders, implantation sites, components, benefits, limitations, and devices associated with various methods of sight restoration, including invasive (retinal prostheses, optic nerve prosthesis, lateral geniculate nucleus prosthesis, and visual cortex prostheses) and non-invasive [transcorneal electrical stimulation (TES), repetitive transorbital alternating current stimulation (rtACS), transpalpebar current stimulation (TpES), low-intensity focused ultrasound (LIFUS), and transcranial random noise stimulation (tRNS)].

Abbreviations: VP or BE: Visual Prostheses or Bionic Eye, RP: Retinitis Pigmentosa, AMD: Age-Related Macular Degeneration, RGC: Retinal Ganglion Cells, DTL: Dacryolingostatoplasty, RPE: Retinal Pigment Epithelium, LGN: Lateral Geniculate Nucleus, TES: Transcorneal Electrical Stimulation, rtACS: Repetitive Transorbital Alternating Current Stimulation, TpES: Transpalpebar Current Stimulation, LIFUS: Low-Intensity Focused Ultrasound, tRNS: Transcranial Random Noise Stimulation.

The placement of a retinal prosthetic may be done is a subject-specific fashion. The electrode array can be placed in epiretinal, subretinal, or suprachoroidal locations. The epiretinal prosthesis is placed directly above the ganglion cells on the inner surface of the retina [e.g., Argus-II Retinal Prosthesis [202] and IMI retinal implant [231]]. It requires an intact inner retinal layer and thus comes with a limitation for extensive retinal degeneration [232, 233, 234]. Notably, surgical complications like retinal damage may occur. The subretinal prosthesis [e.g., Alpha IMS [203] and Prima System implant [204]] is implanted between the retina and RPE. It has a lower risk of retinal damage but a higher risk of retinal detachment [235, 236]. Notably, both epiretinal and subretinal devices are closer to RGCs, thereby providing them direct stimulation that results in higher resolution and specificity [237]. In contrast, the suprachoroidal (STS) implant is placed between the sclera and the choroid, preserving the structure of the retina with lower propensity for surgical complications, such as retinal detachment. However, the indirect stimulation of RGCs leads to a coarser resolution [238]. Moreover, suprachoroidal prostheses have greater stimulation thresholds, raising the possibility of retinal injury [239].

Retinal implants have multiple limitations, mainly causing damage to the retinal structure while planting the electrodes. Additionally, despite technological advancements in retinal implants, the ongoing degeneration throughout the visual pathway may limit the potential of visual prostheses to restore visual perception, because retinal implants are only replacement options [240, 241]. In other words, they do not halt the progression of the ongoing degeneration that may extend from the photoreceptors to the RGCs [94, 242, 243]. A substantial portion of the pathway should still be intact for retinal devices to send decipherable and meaningful electrical impulses. Various engineering challenges in obtaining a high-resolution artificial vision may further limit retinal prosthetics [244, 245, 246].

7.1.2 Optic nerve prosthesis

Optic nerve stimulation is an intriguing method for visual restoration. It is less likely to impair the intricate synaptic processes of the retina, as the electrodes (spiral cuff and penetrating) are implanted far from the retina at the extracranial part of the optic nerve [247]. The spiral cuff electrodes wrap around the optic nerve, thus providing a greater surface area of stimulation, albeit less selectivity, thus requiring a higher number of electrodes for stimulations [248, 249, 250, 251]. The penetrating electrodes, on the other hand, can deliver more focused stimulation [252]. Multiple studies have investigated the direct stimulation of the optic nerve in patients with RP bypassing the damaged retina. Whether using a spiral cuff or penetrating electrodes, the optic nerve stimulation results in the perception of phosphenes across a broad area of the visual field [206, 253, 254, 255, 256, 257, 258].

While optic nerve prostheses hold promise for restoring vision in individuals with optic nerve damage, the technology is still in the experimental stages and faces several challenges, including achieving sufficient visual acuity, ensuring biocompatibility and long-term stability of the device, and developing effective methods for integrating the prosthesis with the brain’s neural circuits. Research in this area is ongoing, with the goal of improving the efficacy and accessibility of optic nerve prostheses for individuals with vision loss.

7.1.3 Lateral geniculate nucleus prosthesis

The LGN is a thalamic structure that receives input from the retina and delivers the visual data to the V1 [259]. It is an interesting site for visual restoration and artificial vision in irreversibly blind patients due to optic neuropathies or retinal degeneration diseases or even in blindness caused by eye trauma [260, 261]. However, it is a more complex surgical procedure but has advantages over other methods. The LGN is a compact structure with a distinct spatial distribution of receptive fields [262], allowing for a less-density electrode array, thus enabling the capture of a broader spectrum of receptive fields. Moreover, anatomically, the LGN is structurally divided into several layers. Magnocellular layers make up the inner two layers (1 and 2), whilst parvocellular layers make up the outer four layers (3, 4, 5, and 6). Additionally, koniocellular layers are filled between them. Since these layers are sensitive to specific features of the incoming visual data (e.g., color or motion perception), they allow more specific stimulation of the downstream layers. Importantly, two separate electrodes in each LGN are needed with precise synchronization for coherent information transfer. Currently, there is an ongoing effort to develop an implant for the LGN, with preliminary results that are very promising [208]. Indeed, results of recent research have confirmed that LGN microstimulation can cause phosphenes and predictable visual percepts [208, 263, 264, 265, 266].

7.1.4 Visual cortex prostheses

As mentioned above, the V1 is the entry point of visual information in the visual cortex and the final option for vision restoration using stimulation [267]. Unlike the LGN, the visual cortex has a large surface area allowing for a higher number of electrodes, thereby enhancing resolution [268, 269, 270]. Visual cortex prostheses bypass damaged areas along most of the visual system pathway, thus providing a potential treatment option for the blind individuals [271, 272] when the retinal/optic nerve/LGN prostheses prove ineffective. Initially, research efforts in the latter half of the twentieth century were aimed at creating artificial vision in blind individuals using surface electrodes, but without any effectiveness for the completely blind individuals [273, 274, 275]. With major advancements in technologies since then, the shift to penetrating electrode arrays has reignited the hope for profoundly blind [276]. Penetrating electrode arrays offer a closer positioning to target neurons with many small electrodes introducing very low currents. This allows offering more precise and discrete phosphene production, but with challenges of biocompatibility and tolerance by the host tissue [277]. Currently, there are active ongoing trials on developing cortical prostheses with enhanced encoding and integration systems [210, 278, 279, 280].

7.2 Non-invasive neurovisual microstimulation techniques

Unlike invasive techniques, non-invasive stimulation not only offers fewer surgical complications but has also been proven to play a neuroprotective role in multiple retinal disorders [281]. These involve introducing an active electrode on the eyelids (transpalpebral or transdermal) or cornea (transcorneal) or around the orbit (periorbital) to deliver electrical stimulation with adjustments for tailoring the parameters (wavelength, frequency, and intensity), contingent upon the individual and their respective phosphene thresholds [282, 283].

7.2.1 Transcorneal electrical stimulation (TES)

It is currently the most frequent type used in retinal diseases [281], involving sterile silver-thread Dawson, Trick, and Litzkow (DTL) electrodes or contact lens as an active electrode placed on the cornea. Although, these electrodes cause fewer corneal complications like abrasion, but require more care for maintaining a proper placement on the cornea [284, 285, 286]. A reference electrode is placed underneath the skin near the active electrode. Multiple clinical trials have shown improved visual function outcomes in patients with different optic and retinal disorders after TcES application [213, 214, 215, 216, 217]. Also, TcES has been reported effective in cases of RP improving visual acuity and field in basic [287, 288] and clinical studies [289]. Additionally, retinal artery occlusions have shown improvement in multifocal electroretinograms (mfERGs) restoring the amplitude of the N1–P1 waves in the ischemic loci [290, 291]. The biological mechanisms by which TcES acts on the neural tissue helping survival and regeneration were mentioned in Section 6.

Interestingly, the effect of TcES is not confined to the retina or optic nerve as it has shown a neuromodulatory effect on the brain by shifting the altered (due to injury/damage) electrophysiological activity in a more normal range for visual [172, 173, 292] function. It is suggested that the visual impairment in blind individuals due to a prechiasmatic lesion is not only related to the primary tissue injury but also to the impaired synchrony and loss/changes of functional connectivity in certain brainwaves (associated with visual perception) [169, 173, 293]. Moreover, prolonged TcES has shown activation of network connectivity and persistent enhancement in the theta, alpha, and beta brainwave synchronization in a retinal degeneration model in both visual and non-visual areas of the brain (V1 and prefrontal cortex) [293, 294, 295]. Xie et al. [296] have found an activation of primary and secondary visual cortices measured by positron emission tomography (PET) and 18 F-fluorodeoxyglucose (FDG) using either DTL or ERG-Jet (contact lens) electrodes. The DTL electrodes resulted in the activation of the V2, while ERG-Jet activated both the V1 and V2, engendering brighter phosphene production.

7.2.2 Repetitive transorbital alternating current stimulation

Repetitive transorbital alternating current stimulation (rtACS) involves the delivery of alternating electrical currents through electrodes placed near supraorbital and infraorbital ridges [297]. This involves a multi-channel device that delivers sinusoidal pulses of oscillatory current through four separate periorbital electrodes bilaterally: two supraorbital and two infraorbital. This method differs from the former in that it makes less contact with the cornea, further lowering the risk of complications like abrasions, dryness, and discomfort [289, 298].

The nature of repetitive stimulation helps in the entrainment of desynchronized brain oscillations, with a periodic application of the externally forced stimulation [299]. Again, the adjustment of the stimulation parameters protocol can be done according to different pathological conditions. It has shown enhanced visual outcomes (visual acuity, visual field, and detection) in optic nerve injury patients [174, 218, 300, 301]. The suggested mechanism of action involves the modulation of oscillatory brain activity, as evidenced by observed alterations in alpha-band brainwave activity, resulting in synchronization [219]. By targeting the visual system through the orbit, rtACS has the potential to modulate neuronal activity and synaptic plasticity in visual pathways. Enhanced neuroplasticity and network connectivity have been shown on EEG along with resynchronization of the alpha bands reported in recent trials [169, 218, 219, 220, 221]. Moreover, Granata and Falsini [302] have observed subjective clinical improvements in a heterogeneous group of various chronic low vision pathologies, associated with an increase in the amplitude of visually evoked potentials (VEP) after the application of rtACS.

7.2.3 Transpalpebral current stimulation

Transpalpebral current stimulation (TpES) involves the application of electrodes on the eyelid, limiting direct contact with the cornea, thereby minimizing the risk of local eye complications like eye discomfort and mucin homeostasis disruption, a lubricating agent on the corneal surface that prevents eye dryness [298]. TpES has the added benefit of lowered IOP, in addition to the similar mechanisms in other methods. In primary open-angle glaucoma (POAG) patients, it acted akin to tyrosine kinase inhibitors and activated maxi-K⁺ channel in the trabecular meshwork (TM) leading to relaxation of TM lowering the resistance to aqueous humor outflow, which in turn lowered the high intraocular pressure [303]. TpES has been proven to be most effective in treating individuals with AMD [222, 223, 224].

7.2.4 Low-intensity focused ultrasound

Low-intensity focused ultrasound (LIFUS) is an emerging neuromodulatory non-invasive method for sight restoration [304, 305, 306, 307]. Numerous studies have consistently shown that FUS effectively stimulates neurons both in laboratory settings, including in-vitro [225, 308], ex-vivo [309, 310], and in-vivo [226, 311] methods. It offers an advantage over other non-invasive techniques like tDCS and TMS by providing a higher spatial resolution (at μm levels) and deeper penetration (at cm levels). It is delivered in low intensity to the nervous system inducing excitatory actions while not causing a neuronal injury [312, 313, 314]. The mechanical energy from the ultrasound waves induces various biological and cellular effects such as cavitation and mechanosensitive ion channel activation [315]. For vision, the LIFUS can help in vision restoration in two ways, either by retinal stimulation or visual cortex stimulation. Recently, retinal non-invasive ultrasound prostheses have been developed for retinal stimulation. However, it has not been approved in clinical settings yet, the basic studies have shown the ability to safely stimulate retinal ganglion cells with different patterns in retinal degeneration models producing phosphenes [225, 226, 227].

An alternative method for using ultrasound in visual restoration involves non-invasively stimulating the visual cortex through transcranial-focused ultrasound. This stimulation leads to temporary modulation of the visual cortex [316, 317, 318, 319]. The parameters of the ultrasound waves have different types of cortical responses. The pulsed waves had a prolonged effect than the continuous waves.

7.2.5 Transcranial random noise stimulation (tRNS)

Another recently emerging non-invasive electrical stimulation technique is transcranial random noise stimulation (tRNS) [320]. A biphasic sinusoidal alternating current with weak intensity (in mA) in unpredictable random frequencies is applied transcranially using two electrodes, either in low-frequency spectrum (<100 Hz) (lf-RNS) or high-frequency spectrum (>100 Hz) (hf-tRNS) [321]. The latter has proven superiority in its neuromodulatory potential [228, 322, 323]. tRNS has the same effect as other non-invasive techniques concerning the modulation of cortical oscillatory activity; however, the exact mechanism is still not well-known [320, 324, 325]. Few studies have shown that the repetitive modulation of voltage-gated sodium channels may explain the results of enhanced neuroplasticity [326, 327, 328]. However, it delivers an alternating electrical current like other techniques such as tACS, and it has proven to be more effective compared with them [329, 330]. This can be explained by the stochastic resonance phenomenon which suggests that delivering noisy stimulation (random frequency current) can amplify the neuron’s sensitivity to weak stimuli [331]. The application of tRNS at the visual cortex promotes visual learning, a rehabilitative approach for functional recovery of central vision loss [332, 333], in both healthy [334, 335, 336] and visually impaired individuals with diseases such as myopia [229], amblyopia [230], and macular degeneration [337].

8. Conclusion

Preclinical research and clinical trials collectively demonstrate that neurostimulation holds promise in enhancing retinal sensitivity, improving visual function, and potentially restoring lost vision. By targeting specific regions of the visual pathway, neurostimulation approaches seek to promote neuronal regeneration and plasticity while avoiding dysfunctional or damaged areas. Despite its potential, neurostimulation faces several challenges and limitations. These include the need for further optimization of stimulation parameters, standardization of protocols, long-term safety, and efficacy assessments, and ensuring broad accessibility of the technology for clinical use. Moreover, the heterogeneity of visual disorders and varied responses to neurostimulation underscore the importance of personalized treatment plans and ongoing research efforts.

Furthermore, researchers are exploring innovative stimulation techniques beyond traditional electrical stimulation, such as optogenetics, magnetic stimulation, and ultrasound stimulation, aiming for more precise and targeted modulation of neural activity within the visual system [338, 339, 340, 341]. Advancements in imaging technologies, such as functional magnetic resonance imaging (fMRI) and optical coherence tomography (OCT), can improve the precise targeting and localization of neurostimulation within the visual pathways, leading to more effective and personalized treatment approaches [342, 343]. Development of closed-loop neurostimulation systems for neuromodulations, which dynamically adjust stimulation parameters based on real-time feedback, may optimize treatment outcomes. Integration of neurostimulation with other therapeutic modalities, such as pharmacotherapy or gene therapy, may synergistically enhance treatment efficacy and promote neuroprotection or neuroregeneration. Continued refinement of non-invasive neurostimulation techniques offers safer and more accessible alternatives for modulating visual cortical excitability. Bridging the gap between preclinical research findings and clinical implementation, along with prioritizing patient-centered outcomes research, is crucial for the widespread adoption of neurostimulation therapies in neuro-ophthalmology and improving the lives of affected individuals.

Contribution and acknowledgements

NS contributed to conceptualization and drafted the original version of manuscript. MK also contributed to drafting the original version of manuscript. SS finalized the figures and contributed to editing. YB, OF, and VB contributed to writing and editing of the manuscript. This research was supported by the University of South Florida (USF) seed grant and the support of the Department of Neurosurgery and Brain Repair, USF.