**1. Introduction**

Thinking of a concept, whether it is a flower or a cat or even a unicorn, can bring up vivid, image-like experiences without external visual input. This is generally referred to as visual imagery or mental imagery, although the latter can extend to other senses (e.g., sound, smell, or touch). The basis of mental imagery has long been debated [1–3] and there is still uncertainty about its neural underpinnings.

Zeman and colleagues [4] gave the inability to generate mental imagery a name, aphantasia, and described individuals with congenital aphantasia who never had this ability. The loss of mental imagery following brain injury—acquired aphantasia—in individuals who had normal imagery before their injury is also well documented, dating back at least to Charcot and Bernard [5] (but see [6]). However, as noted by Farah [7], cases of acquired imagery deficits can be associated with a wide range of lesions (occipital, temporal, or parietal) in either hemisphere, and no other functional deficits

consistently co-occurred with imagery loss with the exception of loss of (visual) dreaming. One plausible reason for this heterogeneity is that mental imagery is not a single phenomenon but can be divided into relatively distinct components, with different underlying anatomy. Some distinguish between a generation process, long-term visual memory, and an inspection process [7], or subsystems such as appearance-based (e.g., shape/color judgment) vs. spatial (e.g., mental navigation/scanning) imagery [8,9] (see also [10]). Supporting this, a meta-analysis of imaging studies showed that while several regions were coactivated during appearance-based and spatial imagery, the former mapped onto the ventral visual stream while the latter evoked specific activity in the dorsal stream [11].

It has been argued that the primary visual cortex (V1) plays a significant role in visual mental imagery [12,13]. Several studies have shown cortical activation in V1 during imagery tasks (e.g., [14–18]) and rTMS (repetitive transcranial magnetic stimulation) targeting V1 can disrupt visual imagery [15]. In addition, individual differences in mental imagery capability covary with differences in V1 surface area [19], V1 functional connectivity [20], and representational overlap between visual imagery and perception in the retinotopic cortex [21]. However, while patients with intact V1 can have severe impairments in mental imagery [22], seemingly intact imagery without a functioning V1 has also been reported [23,24] (see also [25]).

Thus, damage to V1 appears neither necessary nor sufficient for inducing imagery deficits. A review [26] of case studies suggested that extensive left temporal damage is necessary for a visual imagery deficit for object form or color (see also [11]), and more generally that high-level visual areas in the temporal lobe might be particularly important for visual imagery. The fact that patients have been reported to have both high-level visual deficits and selective imagery loss in the same domain (e.g., severe problems in visual recognition and revisualization of faces, [27]), and that actual viewing and visual imagery for particular objects or object categories can evoke a similar pattern of activity in high-level ventral stream regions [21,28,29], is in alignment with the general idea of shared mechanisms between visual imagery and visual perception (for recent reviews, see [30,31]).

Visual imagery and perception however cannot share all mechanisms as there are patients on record with seemingly preserved mental imagery but impaired visual perception [32–36]. For example, case H.J.A. [32] suffered from visual agnosia, achromatopsia, prosopagnosia, alexia without agraphia and topographical impairments. Despite these deficits, H.J.A.'s mental imagery was relatively—albeit not completely—spared. The opposite pattern, impaired visual mental imagery but relatively normal visual perception, has also been reported [37,38]. An example is a patient who had suffered a left occipital and medial temporal infarct. While his visual recognition abilities were generally good, he showed apparent problems in mental imagery such as describing an elephant as having a "tiny waist" and having trouble with verifying sentences that required visual imagery (e.g., "A grapefruit is larger than an orange") [37].

Here we present patient PL518, an architect who reported almost complete loss of visual mental imagery following bilateral stroke in the areas supplied by the posterior cerebral artery (PCA). His responses on the Vividness of Visual Imagery Questionnaire (VVIQ, ad modum [39]) as well as a range of visuoperceptual tests are compared to three other patients with bilateral PCA stroke, as well as another architect with a large unilateral PCA stroke in the right hemisphere. We also compare the structural images of their lesions. The aim of the study is to: (a) describe the correspondence between the perceptual and neuropsychological profile of PL518 compared to the other patients, and (b) to delineate cerebral areas that are uniquely affected in the aphantasic patient and could thus play a fundamental role in the generation of visual imagery.

### **2. Materials and Methods**
