Visual search is an important everyday visual function. Many laboratory studies of visual search use synthetic targets and distractors. Such experiments are somewhat removed from the holistic approach needed to find a face in a crowd, search for an exit sign at an airport or locate a favourite cereal on the supermarket shelf. So we agree with Hulleman & Olivers' (H&O's) contention that “item-based approaches limit the real-world applicability of results from the lab” (sect. 4.1, para. 1). H&O discuss two real-world applications of visual search: radiology and airport security. In this commentary we highlight search as an impaired everyday visual function in people with age-related eye disease. We speculate on how this might be best assessed with the idea of bringing visual search out of the lab and into clinical research, focussing on open angle glaucoma and age-related macular degeneration (AMD), two of the most common causes of visual impairment worldwide (Lamoureux et al. Reference Lamoureux, Pallant, Pesudovs, Tennant, Rees, O'Connor and Keefe2008). Glaucoma is typically associated with peripheral vision loss, whilst AMD causes loss of central vision.

Most studies of visual search in age-related eye disease used an item-based approach (for example, Jacko et al. Reference Jacko, Barreto, Chu, Scott, Rosa and Pappas2000; Reference Jacko, Scott, Barreto, Chu and Fain2001), yet examples taking a more real-world approach are emerging. We investigated visual search in people with glaucoma using two computer-based tasks (Smith et al. Reference Smith, Crabb and Garway-Heath2011), one item-based task requiring participants to identify a target from an array of distractors, and another more real-world task requiring participants to find everyday items in digital photographs of indoor and outdoor scenes. Participants with glaucoma exhibited longer average search times than healthy peers for the real-world task, whilst search times were not significantly different between the two groups for the item-based task. These results support the notion that item-based search tasks are not relatable to real-world applications. A further study (Smith et al. Reference Smith, Glen and Crabb2012), investigating eye movements during the same real-world visual search task, reported a reduction in saccade frequency in people with glaucoma compared with healthy peers. Furthermore, amongst participants with glaucoma, those who made more saccades per second were quicker in finding the real-world targets. These results indicate that eye-movement behaviour is of importance when considering visual search performance of people with age-related eye disease, and were supported by a study of similar design when detecting faces (Glen et al. Reference Glen, Smith and Crabb2013). These findings align with H&O's proposition that fixation count is a critical factor in visual search behaviour.

Fixation count has been investigated in real-world search tasks in AMD. Most visual search research in AMD has been conducted using artificial arrays (for example, searching for a letter T amongst distractors in the form of the letter L) and participants with simulated scotomas (for example Bertera Reference Bertera1988; Coeckelbergh et al. Reference Coeckelbergh, Cornelissen, Brouwer and Kooijman2002; Cornelissen et al. Reference Cornelissen, Bruin and Kooijman2005; Geringswald et al. Reference Geringswald, Baumgartner and Pollman2012; Reference Geringswald, Herbik, Hoffman and Pollman2013; Kuyk et al. Reference Kuyk, Liu and Fuhr2005; MacKeben & Fletcher Reference MacKeben and Fletcher2011; Murphy & Foley-Fisher Reference Murphy and Foley-Fisher1988; Reference Murphy and Foley-Fisher1989). These approaches allow for more controlled experimental design, yet simulated scotomas may not be entirely realistic (Harvey & Walker Reference Harvey and Walker2014; Schuchard et al. Reference Schuchard, Naseer and de Castro1999). One method of simulating central scotoma uses contact lenses with a central opacity, which cause reduced retinal illumination, leading to worsening in visual acuity and contrast sensitivity (Butt et al. Reference Butt, Crossland, West, Orr and Rubin2015). A gaze-contingent simulation of scotoma, incorporating eye tracking, is likely to provide better scotoma simulation (Butt et al. Reference Butt, Crossland, West, Orr and Rubin2015); we have used this in a hazard search task in driving (Glen et al. Reference Glen, Smith and Crabb2015). Results were useful, but simulation cannot capture the real experience of patients, where self-reported perception and description of scotoma varies enormously (Crabb et al. Reference Crabb, Smith, Glen, Burton and Garway-Heath2013). A few studies have investigated real-world visual search in actual patients with AMD; for example Thibaut et al. (Reference Thibaut, Delerue, Boucart and Tran2015) reported individuals with AMD exhibit higher saccade frequencies, shorter fixation durations and longer scan paths compared with those without AMD during visual search. Aspinall et al. (Reference Aspinall, Borooah, Al Alouch, Roe, Laude, Gupta, Gupta, Montarzino and Dhillon2014) found fixation count to be a useful marker of situations subjectively classed as “difficult” by individuals with AMD when assessing eye movement behaviour whilst watching videos of ambulatory journeys. Similarly, Geruschat et al. (Reference Geruschat, Hassan, Turano, Quigley and Congdon2006) investigated gaze behaviour during street crossing and reported higher fixation count during more difficult/visually demanding parts of the task. Seiple et al. (Reference Seiple, Rosen and Garcia2013) observed people with AMD whilst exploring faces and reported fixation count for internal facial features (eyes, nose, and mouth) to be higher for controls than for individuals with AMD. All of these tasks transcend the traditional item-based search.

Studies of everyday visual search have real clinical implications. Visual search in people with visual impairment has been suggested as a predictor for mobility and performance of other daily activities (Kuyk et al. Reference Kuyk, Liu and Fuhr2005). There is evidence for the effectiveness of eye movement training on visual search in congenital prosopagnosia (Schmalzl et al. Reference Schmalzl, Palermo, Green, Brunsdon and Coltheart2008), following brain damage (Bouwmeester et al. Reference Bouwmeester, Heutink and Lucas2007), and for improved visual search and mobility performance in people with visual impairment of ocular origins following repeated practice of an item-based search task (Kuyk et al. Reference Kuyk, Liu, Elliott and Fuhr2010; Liu et al. Reference Liu, Kuyk and Fuhr2007). These types of findings could lead to interventions and alternative approaches to management of patients. Potential also exists for development of tests for detecting and monitoring eye disease by using visual search in both item-based (Loughman et al. Reference Loughman, Davison and Flitcroft2007) and real-world (Crabb et al. Reference Crabb, Smith and Zhu2014) scenarios.

An article published nearly 30 years ago about tumour detection using visual search (Nodine & Kundel Reference Nodine and Kundel1987) states that “detecting an object that is hidden in a natural scene is not the same as detecting an object displayed against a background of random noise.” Research in this area ought to bridge the gap between lab-based testing and the real world. H&O have made an important step towards unifying some of the theory of visual search. We anticipate this will stimulate practical studies that may lead to better understanding of visual search in people with age-related eye disease. In turn we speculate that this will have implications for rehabilitation, and potentially lead to development of new tests for monitoring age-related eye disease.