**6. Visual psychophysical tests in Alzheimer's disease**

To perform psychophysical tests in pathology such as AD can be a great challenge, because most of these tests require understanding and memorisation of the protocol for proper test performance. Therefore, without supervision by an experienced examiner, anomalies detected in testing may actually be the result of a failure to perform the task and not a visual deficit.

#### **6.1. Visual acuity**

neurofibrillary tangles [216, 218, 219] in this nucleus. These pathological changes in the SC may explain the problems of ocular motility frequently found in patients with AD [220].

The suprachiasmatic nucleus (SCN) is a paired structure formed by a group of neurons in the medial hypothalamus, above the optic chiasm, which receives direct RGC input through a retinal‐hypothalamic tract [221]. In addition, it is a primary control centre of circadian rhythms by stimulating melatonin secretion by the pineal gland. It has been seen that there may be marked changes in the SCN in AD, such as decreases in the volume and number of cells (including specific neuronal subpopulations, such as vasopressin and neurotensin neurons) and the formation of neurofibrillary tangles [222–225]. These neuropathological findings may

Pulvinar nuclei, are a collection of nuclei located in the pulvinar thalamus, have reciprocal connections with association areas of the cerebral cortex, in the parietal, occipital and temporal lobes. They consist of several divisions that receive multiple inputs from the visual cortex subdivisions, from the SC and the retina (to a limited extent) [227]. This area is involved in

Numerous amyloid plaques and some neuritic plaques (extracellular deposits within the brain grey matter which are a mixture between amyloid and death neuronal processes) were found along the various subdivisions of the pulvinar nuclei in nine AD patients compared with younger patients and age‐matched controls [228]. This could explain the deterioration of visual

The visual cortex is located in the occipital lobe. It comprises the striate cortex or V1 (primary visual cortex) and visual cortical extrastriate areas such as V2, V3, V4, V5, etc. (secondary visual cortex). Together, the primary and secondary visual cortex consists of a mosaic of several dozen visual areas occupying a large part of the cerebral cortex, approximately 20–25% in humans

In AD, the primary visual cortex is affected after the involvement of other cortical regions [231– 233], except in a variant that manifests with early visual symptoms [234, 235]. Nevertheless, the accumulation of amyloid plaques and neurofibrillary tangles, the decrease in the number of neurons and capillary density, and the reduction of certain enzymes in V1 of AD patients

Although some preliminary studies claimed that there are minimal neurofibrillary tangles and amyloid plaques in primary and secondary visual cortex [237], subsequent studies have found abundant neurofibrillary tangles and plaques in the secondary visual cortex (mostly in the extrastriate 18 and 19 areas) [238, 239]. In both areas 17 and 18, the average neuronal density

be correlated with the misalignment of the circadian rhythm in AD [226].

visual attention and the control of eye movements.

*5.6.2. Suprachiasmatic nucleus*

394 Update on Dementia

*5.6.3. Pulvinar nuclei*

attention in AD patients [174].

are well documented [191, 233–236].

**5.7. Visual cortex**

[229, 230].

The analysis of VA in AD patients was one of the most controversial tests. Although several studies claim that VA is not altered in this neurodegenerative condition [44, 48, 49, 239, 241, 244–246], other researchers find not only a decrease in VA [40, 247] (**Figure 2A**), but they also link this reduction to visual hallucinations when VA is severely decreased [248, 249]. A

**Figure 2.** Mean data of the psychophysical tests. (**a**) Visual acuity, (**b**) contrast sensitivity, (**c**) desaturated Rue 28‐hue colour test, and (**d**) perception digital test. Each bar represents the mean ± SD. \**p* < 0.05 versus control. \*\**p* < 0.01 versus control. Mann‐Whitney *U* test (from **Figure 1** of [40] with permission).

correlation between the worsening in VA and the progression of AD has been described [250], together with a decline in AD patients and in AD patients with vascular dementia under low‐ luminosity conditions [44].

#### **6.2. Colour vision**

To perform the colour test in clinical practice in patients with AD is controversial because these patients have a naming deficit and therefore may have trouble verbalizing colours that they see or they might incorrectly name numbers or shapes. Despite this, some colour‐vision tests do not require verbalisation, although they require concentration and memorisation of the task.

Some published studies which used Farsworth's test and Ishihara's test [251–253] found no differences in colour perception between the AD group and the control one. However, other authors found some tritan‐axis defects, showing a correlation with the degree of dementia [37, 40, 238, 244] (**Figure 2C**). These data agree with the results of other authors [173, 240, 247, 254]. Pacheco et al., in their analysis with Ishihara test and PV‐16, found impaired colour vision consisting of non‐specific mistakes. Such responses were more prevalent in AD patients compared with controls, and were unrelated to the severity of the disease [38]. The disparity in the results may be due to the fact that each study used a different method for studying colour vision, so that comparisons of the results are difficult to interpret. Notably, Salamone et al. claimed that the problem of colour discrimination in AD patients is not purely cognitive but rather seems to be related to the damage of the structures responsible for colour perception [39].

#### **6.3. Visual field**

Like other psychophysical measures, automated perimetry requires considerable cooperation from the patient; therefore the reports on VF and AD are scarce and most are case reports. VF defects in AD vary from homonymous quadrantanopsia [255, 256] to non‐hemianopic VF loss [239]. VF impairment in AD has been found both with manual perimetry [35] and with automated perimetry [34]. The latter showed a significant reduction in differential luminance sensitivity between AD and controls. This study additionally reported that AD patient underwent a diffuse sensitivity loss and, although VF defects involved the central field, deficits were more pronounced in the inferior field, appearing mostly as arcuate defects. They also found that the patients exhibited progressive VF loss 18 months after the initial examination.

#### **6.4. Contrast sensitivity (CS)**

CS tests evaluate the ability of the visual system to discriminate an object from the background in which it is located. This allows us to assess the integration of the information by the RGC and its cortical processing. The CS is measured by a threshold curve in which the spatial frequencies examined are depicted. High spatial frequencies examine the role of parvocellular cells, while low spatial frequencies represent the function of magnocellular cells.

The study of CS in AD has given rise to discrepancies in the results. Most reports have shown that CS function is affected in AD patients, the impairment ranging from a reduction in all spatial frequencies [26, 40, 43, 44, 245, 247, 251, 257–259], to a greater decline in high [26, 40, 258, 260] or low spatial frequencies [239, 244, 261, 262] (**Figure 2B**). By contrast, two studies found no differences between AD patients and controls [48, 246]. Such discrepancies in the results could be due to differences among the patients included in the studies as well as the CS test used [45]. Some CS tests are influenced by VA, such as the Regan chart, a low‐contrast letter, and the Vistech VCTS 6500 whereas others are independent on VA, such as the Pelli‐ Robson test and the Freiburg test [45]. CS impairment in AD patients has consequences for daily functions and cognitive abilities, given that the spatial frequencies most affected appear to be those corresponding to macular function. An example of the importance of CS loss in AD patients is the capacity to predict the risk of falling [45].

#### **6.5. Perception digital test (PDT)**

correlation between the worsening in VA and the progression of AD has been described [250], together with a decline in AD patients and in AD patients with vascular dementia under low‐

To perform the colour test in clinical practice in patients with AD is controversial because these patients have a naming deficit and therefore may have trouble verbalizing colours that they see or they might incorrectly name numbers or shapes. Despite this, some colour‐vision tests do not require verbalisation, although they require concentration and memorisation of the

Some published studies which used Farsworth's test and Ishihara's test [251–253] found no differences in colour perception between the AD group and the control one. However, other authors found some tritan‐axis defects, showing a correlation with the degree of dementia [37, 40, 238, 244] (**Figure 2C**). These data agree with the results of other authors [173, 240, 247, 254]. Pacheco et al., in their analysis with Ishihara test and PV‐16, found impaired colour vision consisting of non‐specific mistakes. Such responses were more prevalent in AD patients compared with controls, and were unrelated to the severity of the disease [38]. The disparity in the results may be due to the fact that each study used a different method for studying colour vision, so that comparisons of the results are difficult to interpret. Notably, Salamone et al. claimed that the problem of colour discrimination in AD patients is not purely cognitive but rather seems to be related to the damage of the structures responsible for colour perception [39].

Like other psychophysical measures, automated perimetry requires considerable cooperation from the patient; therefore the reports on VF and AD are scarce and most are case reports. VF defects in AD vary from homonymous quadrantanopsia [255, 256] to non‐hemianopic VF loss [239]. VF impairment in AD has been found both with manual perimetry [35] and with automated perimetry [34]. The latter showed a significant reduction in differential luminance sensitivity between AD and controls. This study additionally reported that AD patient underwent a diffuse sensitivity loss and, although VF defects involved the central field, deficits were more pronounced in the inferior field, appearing mostly as arcuate defects. They also found that the patients exhibited progressive VF loss 18 months after the initial examination.

CS tests evaluate the ability of the visual system to discriminate an object from the background in which it is located. This allows us to assess the integration of the information by the RGC and its cortical processing. The CS is measured by a threshold curve in which the spatial frequencies examined are depicted. High spatial frequencies examine the role of parvocellular

The study of CS in AD has given rise to discrepancies in the results. Most reports have shown that CS function is affected in AD patients, the impairment ranging from a reduction in all

cells, while low spatial frequencies represent the function of magnocellular cells.

luminosity conditions [44].

**6.2. Colour vision**

396 Update on Dementia

**6.3. Visual field**

**6.4. Contrast sensitivity (CS)**

task.

The PDT is a quick, easy, and sensitive method recently developed for evaluating visual‐ perception disorders in mild AD patients [263]. The test aimed to assess the visual recognition of familiar situations, masked by geometric special effects that hinder perception. Each of the 15 sheets comprising the test shows the same picture at different positions in space. Special effects such as geometric effect (tile) or the effect of the frame 24/48 of MGI Photo Suite III program are used to distort the pictures. The test consists on a set of images that are shown to the patient to identify which one is properly oriented in space. Among the photographs are six common objects, five landscapes, two people, one letter, and one animal. The study of Rami et al. [263] showed that there were significant differences in PDT between mild‐AD patients and control as well as a significant correlation with the MMSE. These results have been recently confirmed by Salobrar‐García et al. [40] indicating that patients with mild AD had significantly more failures than controls and that there was a significant linear association with the MMSE [40] (**Figure 2D**).

#### **6.6. Critical fusion frequency**

The critical fusion frequency (CFF), also called temporal resolution, is a psychophysical threshold and in psychological terms is regarded as a measure of information‐processing capacity [264]. It is defined as the frequency at which an intermittent light stimulus appears to be completely steady to the average human observer [265]. The CFF threshold is determined by the processing in the magnocellular pathway and frontal and parietal cortex [266]. In some studies the CFF appears normal in AD patients [244, 254, 267] with no retinocalcarine abnor‐ mality specific to AD patients [254]. By contrast, other authors found significantly lower CFF and descending scores compared with healthy elderly subjects [239, 264, 268].

#### **6.7. Dark adaptation**

Older adults have serious difficulty seeing under low illumination and at night, even in the absence of ocular disease. This fact can be attributed to delayed rhodopsin regeneration [269]. The study of Rizzo et al. showed that 7 of the 10 AD patients studied had a worse adaptation to darkness than did the control group of slightly younger patients [254].

#### **6.8. Depth perception**

The ability of the human eye to see in three dimensions and judge the distance of an object is called depth perception. Depth perception is grounded in both stereopsis and monocular cues. Because measuring monocular tracks is difficult, studies typically assess stereopsis. When an object is observed, each eye sees it from a slightly different angle. Those images are then sent back to the brain to be integrated into a single image, creating the 3D effect or stereopsis. Stereopsis relies mainly in the primary visual cortex. However, a more detailed analysis reveals that stereoscopic depth takes place in visual‐association areas in the dorsal and ventral cortical pathways [270]. AD patients have been found to have abnormal depth perception in compar‐ ison with controls [238, 251, 254, 271, 272]. Disturbances in stereopsis, motion parallax and interpretation of static monocular cues may result from neuropathology in the AD visual cortex [272]. Other studies investigating stereopsis in AD were inconclusive [267] or found normal operation [171]. More recently, the link between worsening depth perception and AD has been demonstrated by means of functional magnetic resonance imaging (fMRI). The fMRI has revealed hypoactivation in the areas responsible for depth perception [273].

#### **6.9. Motion perception**

Motion perception is the process of deducing the speed and direction of different elements in a scene based on different sensory stimuli: visual, vestibular and proprioceptive. Visual sensory information for motion perception is based on retina [274], LGN [275] and primary and secondary visual cortex processing [276]. There is controversy in the reports on motion perception in AD patients, as some studies have found no differences in this regard [174, 246, 267] while others have identified several deficiencies in the motion perception [173, 277]. Specifically, patients with an early‐stage of AD have great difficulty interpreting the movement of objects, a condition that worsens as the disease progresses [47]. The discrepancy of the results could be related to the test used, since some require discrimination of motion direction while others simply demand the recognition of the occurrence of motion.

#### **6.10. Pupillary response to light**

Pupillary response is controlled by a balance between the cholinergic and adrenergic inner‐ vation being influenced directly or indirectly by the input from the central and autonomic nervous system.

As an impairment of the cholinergic system is known to occur in dementia and AD [278], it would be expected for the pupillary light reflex to be affected in AD [279, 280]. The constriction at the onset of bright light relative to the resting amplitude was significantly reduced in AD compared with healthy age‐matched older adults and young adults [279]. These findings point to the presence of a cholinergic deficit in AD patients [280, 281], as shown by the fact that pharmacological treatment with donepezil, an anticholinesterase agent, partially improves this deficit [280]. However, it is still a challenge to develop a pupillary‐response‐sensitivity test specific to clinical diagnosis of early AD.

In 1994, Scinto et al. reported that AD patients had hypersensitivity of the pupil dilation after diluted tropicamide instillation [282]. Since the publication of this work, numerous studies have been performed, with several results showing both negative [283–291] and positive results [281, 282, 292–297]. Some of these studies point to a lack of specificity to the pupil tropicamide test in AD [288, 298].

#### **6.11. Ocular motility**

**6.8. Depth perception**

398 Update on Dementia

**6.9. Motion perception**

**6.10. Pupillary response to light**

test specific to clinical diagnosis of early AD.

nervous system.

The ability of the human eye to see in three dimensions and judge the distance of an object is called depth perception. Depth perception is grounded in both stereopsis and monocular cues. Because measuring monocular tracks is difficult, studies typically assess stereopsis. When an object is observed, each eye sees it from a slightly different angle. Those images are then sent back to the brain to be integrated into a single image, creating the 3D effect or stereopsis. Stereopsis relies mainly in the primary visual cortex. However, a more detailed analysis reveals that stereoscopic depth takes place in visual‐association areas in the dorsal and ventral cortical pathways [270]. AD patients have been found to have abnormal depth perception in compar‐ ison with controls [238, 251, 254, 271, 272]. Disturbances in stereopsis, motion parallax and interpretation of static monocular cues may result from neuropathology in the AD visual cortex [272]. Other studies investigating stereopsis in AD were inconclusive [267] or found normal operation [171]. More recently, the link between worsening depth perception and AD has been demonstrated by means of functional magnetic resonance imaging (fMRI). The fMRI has

Motion perception is the process of deducing the speed and direction of different elements in a scene based on different sensory stimuli: visual, vestibular and proprioceptive. Visual sensory information for motion perception is based on retina [274], LGN [275] and primary and secondary visual cortex processing [276]. There is controversy in the reports on motion perception in AD patients, as some studies have found no differences in this regard [174, 246, 267] while others have identified several deficiencies in the motion perception [173, 277]. Specifically, patients with an early‐stage of AD have great difficulty interpreting the movement of objects, a condition that worsens as the disease progresses [47]. The discrepancy of the results could be related to the test used, since some require discrimination of motion direction while

Pupillary response is controlled by a balance between the cholinergic and adrenergic inner‐ vation being influenced directly or indirectly by the input from the central and autonomic

As an impairment of the cholinergic system is known to occur in dementia and AD [278], it would be expected for the pupillary light reflex to be affected in AD [279, 280]. The constriction at the onset of bright light relative to the resting amplitude was significantly reduced in AD compared with healthy age‐matched older adults and young adults [279]. These findings point to the presence of a cholinergic deficit in AD patients [280, 281], as shown by the fact that pharmacological treatment with donepezil, an anticholinesterase agent, partially improves this deficit [280]. However, it is still a challenge to develop a pupillary‐response‐sensitivity

revealed hypoactivation in the areas responsible for depth perception [273].

others simply demand the recognition of the occurrence of motion.

For the proper coordination of eye movements the retina, the brain integration of the image, and the extraocular muscles need to be coordinated [299]. It has been reported that in post‐ mortem studies of AD patients, the oculomotor nuclei of the brain are affected [300, 301]. Boxer et al. reported that AD patients displayed reflexive visually guided saccade abnormalities, specifically, prominent increases in horizontal saccade latency [301].

#### **6.12. The electroretinogram (ERG)**

The electroretinogram (ERG) is a record of the bioelectrical response towards light stimuli. The three types used in daily practice are full‐field ERG, pattern ERG (PERG) and multifocal ERG (mfERG). For clinical purposes, full‐field ERG has been standardised so that the results of different researchers can be compared [302].

There is intense controversy with respect ERG impairment in AD patients. Some studies have found that the amplitude and latency of the retinal potentials did not differ between AD and control groups [29, 32, 254, 303, 304], indicating that the PERG may not be valuable in estab‐ lishing an early diagnosis of AD [304]. In the study of Kergoat et al., amplitude and latency of ERG was not affected but there was a delay on the latency of the visual evoked potentials (VEP) [32]. Other studies have reported a significant amplitude reduction in AD patients [25, 26, 305] and postulated that this result is consistent with RGC dysfunction [26].

In PERG examination, increased implicit time of P50‐wave and amplitude reduction in P50‐ and N95‐waves were observed in AD patients [41, 42, 306], and this could reflect an impairment of the magnocellular stream [306]. In addition to PERG, Krasodomska et al. studied pattern VEP in patients with early stages of AD, and normal routine ophthalmological examination results. Their most prevalent findings were amplitude reduction in N95‐wave and increased latency of P100‐wave. Such results showed a dysfunction of RGC and optic nerve in early stages of AD [41]. The mfERG measures macular function [337], and a decrease in electrical activity has been found in the macula of AD patients [307].

#### **6.13. Optical coherence tomography (OCT)**

OCT is a non‐invasive imaging technique that works in a similar way to ultrasound, except that it uses low‐coherence light waves instead of sound waves. The light waves are used to take cross‐section images of the retina. As OCT allows visualisation of the retinal layers, their thickness can be mapped and measured. In fact, this technique has already been established as the standard image model for retinal tests (**Figure 1**).

Currently, RNFL, RGC and inner layers of the retina are considered indirect biomarkers of the CNS, enabling the prediction of brain pathology in patients suffering from different neuro‐ logical diseases [184, 308]. Since the development of OCT, this tool has been used to measure the thickness of the RNFL in different neurodegenerative pathologies. Despite that OCT was first developed in 1991 and commercially distributed in 1995, it was not until 2001 when a study was first published on the thickness of the RNFL in patients with AD [42].

During those first years, many studies appeared focusing on the peripapillary RNFL thickness in AD patients. In every study, a significant decrease in RNFL thickness was objectified in AD patients and compared with age‐matched controls. These analyses were carried out by segmenting the measures of the peripapillary thickness according to the area (superior, inferior, nasal and temporal). Several studies showed a decrease in the peripapillary RNFL thickness in all areas [10, 14, 16, 18, 42, 309]; others found that the thinning occurred in the inferior and superior regions [17, 307, 310], while in still other studies this significant decrease appeared only in the superior peripapillary region [15, 213, 311–313]. Some studies reported a certain thinning in the RNFL associated with the progressive cognitive decline [203, 312, 314]. Some authors even suggest that the inferior peripapillary quadrants might be the area with most specificity and sensitivity regarding the detection of the cognitive decline in the initial stages of the AD [17]. However, Salobrar‐García et al. [20] reported that their group of patients with mild AD showed no significant difference with respect to control subjects in terms of RNFL thickness of the peripapillary region (**Figure 3A**). These authors postulated that although no statistically significant differences in peripapillary RNFL were found between control and AD eyes, the increase in peripapillary thickness observed in mild‐AD patients could be secondary to an inflammatory process that may represent an early stage of degener‐ ation and could lead to progressive peripapillary fibre damage. The variability in peripapillary RNFL thickness reported in AD might be due to differences in disease progression among patients studied, since patients with greater involvement of the peripapillary region were those with a more advanced stage of AD.

**Figure 3.** Mean data of RNFL thickness against eye quadrants assessed with optical coherence tomography (OCT). (**a**) Peripapillary segmentation retinal nerve fibre layer, (**b**) central macular ring (1 mm away from the fovea). \**p* value < 0.01 (modified from **Figure 2** [20] with permission).

Recently, some studies focusing on the analysis of patients with mild cognitive impairment (MCI) found a thinning of the peripapillary region [62, 314–316]. MCI patients also have a thinning in the macular ganglion cell‐inner plexiform layer [317]. By contrast, Ascaso et al., reported an increase in the macular volume in MCI when compared to control subjects and AD patients [315]. These findings could be explained as an increase in the macular volume caused by a possible inflammation and gliosis prior to neuronal cell death.

Recently, studies in AD analysed the measurement of not only the peripapillary RNFL thickness but also the RNFL thickness in the macular region. They demonstrated a significant RNFL thinning in the macular region of AD patients compared with age‐matched controls [14, 19, 62, 63, 307, 311, 315] (**Figure 3B**). A study carried out using the latest OCT technology, which allows an analysis of the different retinal layers separately, noted that the thickness reduction occurred in the inner layers of the retina (RNFL‐RGC complex), whereas the outer layers were not affected [63, 203].

In the most incipient AD stages (mild AD), psychophysical tests having the greatest predictive value are reportedly the CS, VA, unspecific errors in tritan region and the PDT [40]. In addition, the macular RNFL thickness and total macular volume measured by OCT have highly significant sensitivity and specificity for differentiating mild AD patients from healthy subjects, the thickness of the inner upper macular RNFL seeming to have the highest diagnostic value in mild AD neurodegeneration. Probably, the first affected area of the retina in mild AD is the macular area, where, due to the arrangement of the multilayer bodies of the ganglion cells, the decrease is easier to detect [19, 20]. These observations highlight the importance of applying psychophysical tests and OCT in patients with incipient AD stages.

Due to the ageing population increasing, the incidence of neurodegenerative diseases such as AD is growing. As demonstrated by the results of the visual psychophysical test mentioned above, the eye gives us a valuable window for evaluating these neurodegenerations. Therefore, the inclusion of ophthalmological examination could become an important tool in early diagnosis and follow‐up of these patients.
