Seeing Through Psychosis

Ineluctable modality of the visible, at least that if not more, thought through my eyes.

—James Joyce, Ulysses

We do not see the world as it is. Rather, our visual experience is the end-product of a neurobiological pathway of staggering complexity: photons activate rods and cones; the resulting signal is processed numerous times before it even leaves the retina; it then continues through the lateral geniculate nucleus, to the primary and secondary visual cortices, and on to association cortices before it is ultimately incorporated into conscious awareness. Across this pathway, the visual system extracts basic components (e.g., shape, color, speed, location), processes each of them separately, reassembles them into a complete picture, collates them with past experiences, and may even insert what it expects to find. The final product can be thought of as the ultimate, personal “virtual reality.”

With all of this complexity, an obvious question looms: What happens when the system malfunctions? What if an individual begins experiencing the world differently from others? This idea is both philosophical and entirely practical. It is one of the oldest questions in psychiatry: What does it mean for a person to lose their connection to reality?

Emil Kraepelin, the 19th century German psychiatrist, was one of many great minds to wrestle with this topic. Believing in the power of observation, he took detailed accounts of patients whose behavior differed from societal norms. From this work, he developed the construct of Dementia Praecox. But he wanted to delve deeper—to understand not only their outwardly observable characteristics, but also their internal experience. This curiosity led him to a seemingly minute but profound discovery: that individuals with dementia praecox demonstrate incomplete perception of very briefly exposed stimuli (1). In an alternate timeline, this could have launched an exciting research movement exploring how patients with psychosis see the world.

Unfortunately, over the next several decades, these findings were lost in the annals of psychiatry. One researcher after another focused on more dramatic elements of psychosis. Eugen Bleuler expanded on Kraepelin’s description of cognitive deficits in patients with schizophrenia but concluded that both perceptual and sensory functioning were intact (1). Kurt Schneider developed the concept of first-rank symptoms, thereby shifting attention to auditory hallucinations and bizarre delusions. The discovery of chlorpromazine in the early 1950s further solidified focus on these “positive symptoms,” which were most disturbing to others and, for the first time, treatable. The reliance on readily—and largely outwardly—observable attributes was further concretized with the publication of the DSM-III. And so Kraepelin’s observations on sensory deficits became obscured in the darkness of history.

But the wheels of science were in full motion. By the mid-late 20th century, the new field of cognitive neuroscience was emerging—combining decades of psychological research with novel approaches in neuroscience and with insights from a range of traditionally distinct disciplines (e.g., linguistics, computer science, and anthropology). Key contributions came from earlier work in psychophysics—a branch of science dedicated to the systematic and quantitative study of perception. A central goal of cognitive neuroscience was to understand how the brain receives, processes, and transforms information. In short, how do we make sense of the world?

In the late 1980s, researchers began applying psychophysics approaches to clinical populations, including individuals with schizophrenia. As they did so, they discovered a range of abnormalities that built on Kraepelin’s early observations: both in basic visual functions (e.g., perception of contrast, speed, and orientation) and also in higher-level processes (e.g., visual context integration and visuospatial working memory) (2). It turned out that the differences were more wide-reaching than Kraepelin had thought.

Moreover, the differences in visual perception were not just academic, isolated findings; they were directly associated with clinical symptoms of schizophrenia. For example, Peter Uhlhaas and colleagues demonstrated that impairments in the ability to group the pieces of an image into a coherent whole (perceptual organization) were associated with disorganization symptoms (3). Other research found that deficits in motion processing and visuospatial memory were associated with negative symptoms (4). More broadly, visual impairments have been correlated with lower overall functioning and poorer treatment response in patients with schizophrenia (2).

As more data accumulated, researchers wondered: Do visual impairments occur only in the context of overt clinical pathology or might they be heritable traits that predate clinical symptoms or exist, more broadly, as a psychosis-related phenotype?

A range of data have suggested the latter. Visual perceptual impairments have been found in clinical high-risk groups for psychosis, first-degree relatives of individuals with psychosis, and even healthy individuals with schizotypal traits (2). This indicates that specific visual impairments may reflect “endo-phenotypes” of psychosis (i.e., measurable biological characteristics that are associated with an illness due to shared genetic influences). Consistent with this idea, studies have demonstrated a link between visual abnormalities in childhood and adolescence and later development of schizophrenia (5). Together, these studies illustrate how careful assessment of visual perception could offer a critical tool for identifying those individuals at highest risk.

At the same time, these findings raised an even more provocative question: Might visual deficits somehow contribute to the formation of psychotic symptoms?

Pamela Butler and colleagues were fascinated by this question. More specifically, they wondered whether differences in basic visual processing could contribute to downstream symptomatology. In an elegant set of experiments, they found that when they presented subjects with pictures of faces at varying levels of contrast, individuals with schizophrenia showed marked impairments in their ability to recognize emotions (6). Moreover, the data suggested that specific deficits in the magnocellular pathway (critical for the rapid recognition of objects) caused deficits in contrast perception; these, in turn, contributed to the impairments in emotion recognition. The real-world significance of such a deficit is intuitive: having trouble distinguishing stimuli (especially in low-contrast settings, like a dimly lit alley) could cause difficulty recognizing emotions on faces; this could set the stage for misinterpreting social cues and thereby contribute to paranoia.

Butler’s paper illustrated one way in which basic visual impairments could have bottom-up effects on higher-order processes in psychosis. Several other examples support this conceptual framework. Kim et al. (7) have shown that differences in basic motion processing may lead to difficulty processing biological motion, thereby causing problems understanding other people’s actions and intentions. As above, this effect might be most significant in real-life situations with ambiguous signal (in this case when there are multiple sources of motion, as in crowds). Other studies have suggested that early visual processing deficits may contribute to various impairments in higher cognitive functions such as working memory and attention (2). Together, these studies support the radical idea that visual perceptual deficits may play a causative role in forming some elements of the psychosis syndrome. This work dovetails with a separate, emerging field of inquiry that explores from a computational perspective how prior experiences (across perceptual domains) may create expectations that shape perception through feed-forward mechanisms (8).

If true, these findings also offer a tantalizing possibility: can we treat visual deficits before psychotic symptoms emerge and potentially mitigate the development of psychopathology?

Although limited in number, early studies in individuals with psychosis indicate that visual perceptual training may improve deficits in visual processing and may even improve cognitive functioning (9). This work is still in its infancy and more data are necessary to determine whether such interventions are clinically meaningful in preventing the onset of psychosis in high-risk populations. However, given their low risk profile compared to other interventions (such as dopamine-blocking medications), visual trainings offer a compelling avenue to explore.

One major challenge to realizing the potential of this research is the insidious nature of the problem: because deficits are often subtle, they may not be “visible” to either patients or physicians. Moreover, their possible impact on clinical symptoms may be overlooked. For example, when an individual with psychosis has trouble reading, their problem may be attributed to attention deficits, but reading problems may also result from impaired contrast perception (10). Another obstacle to capitalizing on this research is that visual processing cannot yet be readily assessed in clinical settings. Efforts to develop reliable, validated, and easily implementable tools are now underway, with promising early results (4).

For more than a century, psychiatrists have struggled to make sense of a complex illness that emerges from our most complex organ. Over this time, the central limiting factor has been the lack of reliable tools to connect phenomenology to differences in brain function. Cognitive neuroscience—and, more specifically, vision science—offers a rare experimental opportunity: to study a system with well-described anatomical and functional details using objective, validated measurements. This is one approach of many—similar work is ongoing for auditory processes. Capitalizing on basic perception science will enable us to better understand how our patients experience the world and potentially help them see it more clearly.