Differential brain activity in visuo-perceptual regions during landmark-based navigation in young and healthy older adults

Older adults exhibit prominent impairments in their capacity to navigate, reorient in unfamiliar environments or update their path when faced with obstacles. This decline in navigational capabilities has traditionally been ascribed to memory impairments and dysexecutive function whereas the impact of visual aging has often been overlooked. The ability to perceive visuo-spatial information such as salient landmarks is essential to navigate in space efficiently. To date, the functional and neurobiological factors underpinning landmark processing in aging remain insufficiently characterized. To address this issue, this study used functional magnetic resonance imaging (fMRI) to investigate the brain activity associated with landmark-based navigation in young and healthy older participants. Twenty-five young adults (μ=25.4 years, σ=4.7; 7F) and twenty-one older adults (μ=73.0 years, σ=3.9; 10F) performed a virtual navigation task in the scanner in which they could only orient using salient landmarks. The underlying whole-brain patterns of activity as well as the functional roles of scene-selective regions, the parahippocampal place area (PPA), the occipital place area (OPA), and the retrosplenial cortex (RSC) were analyzed. We found that older adults’ navigational abilities were diminished compared to young adults’ and that the two age groups relied on distinct navigational strategies to solve the task. Better performance during landmark-based navigation was found to be associated with increased neural activity in an extended neural network comprising several cortical and cerebellar regions. Direct comparisons between age groups further revealed that young participants had enhanced anterior temporal activity. In addition, young adults only were found to recruit occipital areas corresponding to the cortical projection of the central visual field during landmark-based navigation. The region-of-interest analysis revealed increased OPA activation in older adult participants. There were no significant between-group differences in PPA and RSC activations. These results hint at the possibility that aging diminishes fine-grained information processing in occipital and temporal regions thus hindering the capacity to use landmarks adequately for navigation. This work helps towards a better comprehension of the neural dynamics subtending landmark-based navigation and it provides new insights on the impact of age-related visuo-spatial processing changes on navigation capabilities.

subtended approximately 34 x 20 degrees of visual angle. 165 The virtual environment was programmed with Unity3D game engine (Unity 166 Technologies SF; San Francisco, CA; https://unity.com/) and had participants navigate 167 actively in a first-person perspective. The virtual environment was a three-arm maze (Y-168 maze) consisting of three corridors radiating out from a center delimited by homogenous 169 wooden-like walls. Two configurations were designed. In the landmark condition all arms 170 1 All older participants scored 28 or above on the MMSE except one participant who scored 24. We decided to include this subject nonetheless as his extended neuropsychological evaluation was normal and no significant changes were detected when removing him from the fMRI analyses were 18 virtual meters (vm) long and equiangular. Three light gray-colored objects (a square, 171 a triangle and a circle) were placed in front of each short wall at the center of the maze 172 (Figure 1-A). In the control condition, the arms were 18 vm long and equiangular, and the 173 maze was devoid of objects (Figure 1-B). 174 Participants navigated actively through the virtual environment with an MRI-175 compatible ergonomic two-grip response device (NordicNeuroLab, Bergen, Norway). They 176 could move forward (thumb press), turn right (right index press) and turn left (left index 177 press). A single finger press was necessary to initiate or stop movement. The forward speed of 178 movement was set at 3 vm/s and the turning speed at 40°/s.  The scanning session during the navigation task was divided into three runs: an 185 encoding phase and a retrieval phase for the landmark condition and a control condition. At 186 the start of the encoding phase participants were positioned in the center of the maze 187 randomly facing one of the three arms. They were instructed to find a goal (gifts) hidden at 188 the end of one corridor and remember its location using the visual information available in the 189 center of the environment (the three light gray-colored objects). The encoding phase lasted 3 190 min to ensure that participants could explore all corridors. The retrieval phase in the same 191 environment then began. In each trial participants were placed at the end of one of the two 192 corridors that didn't contain the goal with their back against the wall. The starting positions 193 across trials were pseudo-randomized across subjects. Participants were asked to navigate to 194 the previously encoded goal location. Upon arrival at the end of the correct arm, the gifts 195 appeared to indicate successful completion of the trial and a fixation cross on a gray screen 196 was presented for an inter-trial interval of 3-8 s. Participants needed to complete seven trials. 197 The control condition consisted of a retrieval phase only; it was designed to account for 198 potential confounding factors such as motor and simple perceptual aspects of the task. It was 199 always performed last and it comprised four trials. Subjects started from the end of an arm 200 and moved to the center of the maze from where the target was readily visible. They were 201 instructed to navigate towards it. For both conditions, we recorded the trial duration and the 202 response device use. 203 A short debriefing phase concluded the experimental session. Participants were probed 204 on the strategy they used to orient in the landmark condition. They were asked to report how 205 they solved the task: i) using one object, ii) using at least two objects, iii) randomly, iv) other 206 strategy. Participants were deemed to be using a place-based strategy when their decision was 207 based on two landmarks or more and to be using a response-based strategy when their  blocks of objects, 2 blocks of scrambled objects, and 4 blocks of fixation). Each stimulus was 220 presented for 400 ms followed by a 600 ms inter-stimulus interval. Participants performed a 221 "one-back" repetition detection task.    The first five functional volumes of the encoding, retrieval and control runs were discarded to 251 allow for equilibration effects. Slice-timing correction was applied and functional images 252 were realigned to the mean functional image using a rigid body transformation. Artefacts 253 related to motion were then examined with ArtRepair. Two older subjects were subsequently   Time series for each voxel were high-pass-filtered (1/128 Hz cutoff) to remove low-frequency 271 noise and signal drift. Individual contrasts were submitted to a multiple regression and a two-272 samples t-test. Sex and total brain volume were included as covariates in the regression and 273 total brain volume was included as a covariate in the two-samples t-test (see section 3.1).

274
Areas of activation were tested for significance using a statistical threshold of p < 0.001     Table 1.

304
Navigation performance across age groups is presented in Figure 2. We found that 305 older adults chose the wrong corridor in 10% of trials while young adults made no errors (X²

306
(1, N = 42) = 12.35, p = 0.006; Figure 2A). In addition, older subjects were significantly 307 slower to reach the goal in the landmark condition than younger subjects (mean ± SEM: 19.85 308 s ± 1.67 vs 11.97 s ± 0.13; U(40) = 5.267, p < 0.001; Figure 2B). There was no sex effect on 309 navigation time, defined by the average time to reach the goal, in each age group separately.

310
However, when data from both age groups were pooled women's navigation time appeared to 311 be significantly longer than men's (17.75 s ± 1.88 s vs 13.40 s ± 0.64 s; U(40) = 2.294, p = 312 0.022). Sex and total intracranial volume were therefore included as covariates in the fMRI 313 multiple regression analyses. We further found that age was a significant predictor of strategy 314 use (R 2 = 0.967, p = 0.048). Older adults were less likely to rely on a place-based strategy 315 during landmark-based navigation than younger adults ( Figure 2C). To locate the brain regions related to navigation performance, we first examined the 319 association between both groups' navigation time and patterns of brain activity for the fMRI 320 contrast [Landmark > Control]. We observed a negative association between navigation time 321 and neural activity in many clusters across the brain (Table 2 and  reported is largely driven by age. 334 We then investigated the relationship between navigation time and neural activity in 335 each age group separately (Table 2). In the young participant group, navigation performance

Two-sample analyses 341
Results for the within-group and between-group analyses are shown in Table 3 and  (Table 4). These results hint at the possibility that the control 359 condition was cognitively more demanding for older participants than for young participants.

375
The present fMRI study examined age-related differences in landmark-based 376 navigation using a non-ambiguous Y-maze reorientation paradigm. The task was designed to 377 limit the influence of mnemonic and motor components in order to gain a specific 378 understanding of the neural bases subtending visual spatial cue reliance in young and healthy 379 older adults.    495 Surprisingly, we did not find differences in the activity of the RSC across age groups.

535
To conclude, the present study shed light on the possibility that navigational deficits in 536 old age are linked to functional differences in brain areas involved in visual processing and to 537 impaired representations of landmarks in temporal regions. This work helps towards a better 538 comprehension of the neural dynamics subtending landmark-based navigation and it provides 539 new insights on the impact of age-related spatial processing changes on navigation 540 capabilities. We argue that approaching the study of spatial navigation in healthy and 541 pathological aging from the perspective of visuo-perceptual abilities is a critical next step in     by between-group analyses (total intracranial volume was included as covariate). The