The oblique effect: The relationship between profiles of visuospatial preference, cognition, and brain connectomics in older adults

Highlights

• Oblique effect = advantage for stimuli angled horizontal/vertical rather than diagonal.

• Older non-demented adults produced more oblique than horizontal/vertical (HV) errors.

• Lower executive functioning was associated with a larger oblique effect.

• Differential patterns of neural connectivity associated with oblique and HV errors.

Abstract

The oblique effect (OE) describes the visuospatial advantage for identifying stimuli oriented horizontally or vertically rather than diagonally; little is known about brain aging and the OE. We investigated this relationship using the Judgment of Line Orientation (JLO) in 107 older adults (∼age = 67.8 ± 6.6; 51% female) together with neuropsychological tests of executive functioning (EF), attention/information processing (AIP), and neuroimaging. Only JLO lines falling between 36-54° or 126–144° were considered oblique. To quantify the oblique effect, we calculated z-scores for oblique errors (zOblique = #oblique errors/#oblique lines), and similarly, horizontal + vertical line errors (zHV), and a composite measure of oblique relative to HV errors (zOE). Composite z-scores of EF and AIP reflected domains associated with JLO performance. Graph theory analysis integrated T1-derived volumetry and diffusion MRI-derived white matter tractography into connectivity matrices analyzed for select network properties. Participants produced more zOblique than zHV errors (p < 0.001). Age was not associated with zOE adjusting for sex, education, and MMSE. Similarly adjusted linear regression models revealed that lower EF was associated with a larger oblique effect (p < 0.001). Modular analyses of neural connectivity revealed a differential patterns of network affiliation that varied by high versus low group status determined via median split of zOblique and zHV errors, separately. Older adults exhibit the oblique effect and it is associated with specific cognitive processes and regional brain networks that may facilitate future investigations of visuospatial preference in aging.

Introduction

The oblique effect represents a “consistent superiority in performance when visual stimuli are horizontal or vertical, as opposed to oblique” or on a diagonal (Appelle, 1972). Ophthalmological, neurophysiological, and psychological literature suggest that the oblique effect is consistently observed in humans (Arakawa et al., 2000; Attneave and Olson, 1967; Essock, 1980; Higgins and Stultz, 1950; Maffei and Campbell, 1970; McMahon and Macleod, 2003; Orban et al., 1984; Westheimer, 2003) and multiple animal species (De Valois, Yund and Hepler, 1982; Geisler and Albrecht, 1997; Li et al., 2003), including non-human primates (De Valois et al., 1982; Mansfield, 1974). The oblique effect has been noted as early as three months of age in humans (Leehey et al., 1975; Sokol et al., 1987), suggesting this phenomenon is innate, rather than learned. Despite the many changes in brain structure and cognition that occur with age, less is known about how aging and age-related cognitive profiles are associated with the oblique effect.

To our knowledge, only two studies have examined the oblique effect in older adults and only one of which reported results specific to normal cognitive aging. While mid-to late-life healthy control participants tended to mis-identify one oblique line for another, the oblique effect, defined as greater errors during oblique relative to non-oblique (i.e., horizontal or vertical) line identification, was not explicitly investigated (Ska et al., 1990). Meanwhile, participants with dementia made more oblique line and non-oblique line errors indiscriminately when compared to controls (Ska et al., 1990) suggesting that patients with probable Alzheimer's dementia showed visuospatial impairment characteristic of global judgment errors on line orientation. Finton and colleagues (Finton et al., 1998) further confirmed that when compared to healthy controls, participants with Alzheimer's dementia failed to display any specific error phenotype. Some researchers hypothesize that subtle cognitive alterations in visuospatial functions may be an early indicator for the development of dementia (Cabeza et al., 2004; Dolcos et al., 2002; Lamar et al., 2016), as opposed to emerging as a global phenotype seen at later stages of disease (Finton et al., 1998; Ska et al., 1990). Thus, investigating for the presence of the oblique effect and how this error profile is potentially altered by age, and age-associated cognition and/or brain structure, may help to identify early cognitive markers of neurodegenerative disease in vulnerable individuals.

In the past two decades, neurophysiological and neuroimaging research on the oblique effect has provided converging evidence of distinct regional brain involvement in the divergent line orientations that contribute to this form of visuospatial advantage. For example, animal studies investigating the oblique effect report that neurons in the striate cortex respond more frequently to horizontal and vertical stimuli than oblique stimuli (Bonds, 1982; Coppola et al., 1998; De Valois et al., 1982; Dragoi et al., 2000; Li et al., 2003). Furthermore, studies of visual evoked response potentials in humans demonstrate a more robust response in the striate cortex to horizontal and vertical stimuli than they do to oblique stimuli (Arakawa et al., 2000; Maffei and Campbell, 1970; Sokol et al., 1987). Neuroimaging results using functional magnetic resonance imaging (fMRI) also report increased blood oxygenation level dependent (BOLD) responses within the striate cortex during the presentation of horizontal and vertical task items compared to oblique-oriented items (Furmanski and Engel, 2000). Little work exists attempting to extend these results – specifically, higher striate involvement when viewing horizontal and vertical versus oblique-oriented lines – to feedforward connections, i.e., extrastriate areas or other cortical regions, particularly as it relates to oblique effect error profiles in humans. The research that does exist comes from lesion studies of overall performance on judgments of line orientation suggesting right (greater than left) posterior involvement (Benton et al., 1975; Mehta and Newcome, 1991). Recent advances in neuroimaging may provide an in-depth opportunity to examine the organizational networks that mediate the oblique effect in healthy older adults.

Connectomics allows for a graph-theoretical assessment of system properties in order to understand quantitatively how brain regions, or ‘nodes’, communicate and interact (Rubinov and Sporns, 2010). Additionally, advanced graph-theoretical ‘modularity analysis’ investigates how a group of nodes preferentially interact among themselves to form a community or module, which can then be compared between groups of brain networks to assess for ‘modular’ differences (GadElkarim et al., 2012; Ye et al., 2015). Understanding the age-related associates of oblique effect error profiles, as well as the brain connectome neurocircuitry underpinning it, may enhance our knowledge of this long-studied visuospatial phenomenon of in older adults given the paucity of work conducted in older adults to date (Finton et al., 1998; Ska et al., 1990), and foster future work investigating subtle visuospatial markers of pathological aging.

The present study focused on the relationships between aging, cognition, neural connectivity, and the oblique effect in 107 non-demented, non-depressed older adults. We first hypothesized that 1a) older adults would make more errors on oblique lines compared to horizontal and vertical (HV) lines after controlling for relevant confounders; and, 1b) age would be positively associated with the oblique effect after controlling for relevant confounders. Previous research has documented the importance of executive functioning and, to a lesser degree, attention/information processing to the alterations seen in visuospatial processing in older adults regardless of dementia (Freeman et al., 2000; Lamar et al., 1997). Thus, our second hypothesis was that lower performance on an executive function composite would be associated with a larger oblique effect, i.e., a higher number of oblique relative to HV errors, after controlling for relevant confounders. Given that older age and lower performance on executive function tasks have both been associated with reductions in white matter integrity and altered structural connectivity in otherwise healthy older adults (Charlton et al., 2006; Charlton et al., 2010; Gonzales et al., 2017; Lamar et al., 2016), our third and final hypothesis was that a larger oblique effect would be associated with lower graph-theoretical metrics of neural connectivity and that higher levels of oblique errors but not HV errors would be associated with distinct modularity within anterior (i.e., prefrontal) regions of brain.