Functional Neural Correlates of a Useful Field of View (UFOV)-Based fMRI Task in Older Adults

Abstract

Declines in processing speed performance occur in aging and are a critical marker of functional independence in older adults. Studies suggest that Useful Field of View (UFOV) training may ameliorate cognitive decline. Despite its efficacy, little is known about the neural correlates of this task. Within the current study, 233 healthy older adults completed a UFOV-based task while undergoing functional magnetic resonance imaging (fMRI). During the “stimulus” portion of this task, participants must identify a target in the center of the screen and the location of a target in the periphery, among distractors. During the “probe” portion, participants must decide if the object in the center and the location of the target in the periphery were identical to the “stimulus” screen. Widespread bilateral whole-brain activation was observed when activation patterns of the “probe” contrast were subtracted from the “stimulus” contrast. Conversely, the subtraction of “stimulus” from “probe” was associated with discrete activation patterns consisting of 13 clusters. Additionally, when evaluating the variance associated with task accuracy, specific subregions were identified that may be critical for task performance. Our data elucidate the functional neural correlates of a UFOV-based task, a task used in both cognitive training paradigms and assessment of function.

Introduction

It is well established that cognitive abilities, including memory, processing speed, attention, and executive functioning decline with increasing age, even in the absence of pathology. These age-related cognitive declines, known as cognitive aging, are associated not only with diminished cognition, but with poorer quality of life, health status, driving outcomes and disruptions in complex activities of daily living (Harada, Love and Triebel 2013; Miller et al. 2013; Feng et al. 2017; Vance et al. 2007). However, prior studies have demonstrated robust associations between a variety of cognitively stimulating activities and decreased risk of dementia and mild cognitive impairment (MCI; Wilson et al. 2002; Verghese et al. 2003; Marioni et al. 2012). These findings suggest that the cognitive aging trajectory is malleable, and able to be altered through external interventions. Given the ever-growing older adult population, there is a profound need for cognitive interventions in slowing cognitive aging trajectories.

Efforts have been made to ameliorate cognitive declines with the application of standardized cognitive training. Randomized controlled trials have demonstrated efficacy of domain-specific cognitive training on enhancing processing speed performance (Ball et al. 2002; Vance et al. 2007; Simpson et al. 2012; Wolinsky et al. 2013). One particularly efficacious cognitive training paradigm is Useful Field of View (UFOV) cognitive training. Studies using UFOV cognitive training interventions demonstrated cognitive gains lasting 10 years following study completion, with each UFOV training session associated with up to a 29% reduction in dementia prevalence (Ball et al. 2002; Rebok et al. 2014; Edwards et al. 2017). UFOV cognitive training performance has been associated with better driving outcomes (Roenker et al. 2003; Ball et al. 2010; Fausto et al. 2021) and improved health-related quality of life (Wolinsky et al. 2009, 2013). Additionally, UFOV test performance has been associated with fewer difficulties in completing timed IADLs (Owsley et al. 2001; Valdés et al. 2017).

The UFOV cognitive training paradigm is based on a previously validated visual assessment battery, referred to as the UFOV test. The traditional UFOV test is comprised of 4 subtasks with increasing difficulty, with scores reflecting the briefest presentation time in which participants can accurately respond. The first subtask requires participants to identify a central target (either a car or truck) presented in the center of screen. The second subtask is similar to the first, with the addition of a peripheral target. In this subtask, the participant must identify which central target was presented and the radial location of the peripheral target. The third subtask is identical to the second, with the addition of distractors among the peripheral target. The fourth subtask includes the addition of a second central target. The UFOV test has been widely validated, and has demonstrated high test–retest reliability (Ball and Owsley 1993; Ball et al. 2002; Edwards et al. 2005; Edwards et al. 2018).

Despite the UFOV cognitive training paradigm’s efficacy in improving cognitive and functional outcomes, surprisingly few studies have investigated the neural correlates of the UFOV test and its subtasks. Only 2 small studies to date have assessed functional blood-oxygen level dependent (BOLD) activation patterns relating to a task similar to third subtask of the UFOV test. Scalf et al. (2007) demonstrated that behavioral performance was related to activation in the right inferior frontal gyrus and the right ventral precentral gyrus. Additionally, a study by Ross et al. (2018) found that a UFOV-based intervention including UFOV subtask 3 was associated with within-subjects BOLD activation changes in the anterior cingulate cortex (ACC), anterior insula, dorsolateral prefrontal cortex (DLPFC), inferior parietal lobule (IPL), supplementary motor area (SMA) and thalamus on an in-scanner adaption of UFOV subtask 2. To date, no studies have directly evaluated the neural underpinnings of UFOV subtask 3 during functional magnetic resonance imaging (fMRI). As UFOV subtask 3 reflects the critical components currently deployed in commonly used UFOV cognitive training paradigms (e.g., Double Decision from PositScience BrainHQ), there is a pressing need to better understand its neural correlates.

As such, the current study aimed to elucidate the mechanistic neural underpinnings of the UFOV test using a modified UFOV task specifically adapted for fMRI setting. This adapted task, henceforth called the UFOV fMRI task, is based on subtask 3 of the UFOV test, with the inclusion of a central target and a peripheral target among distractors. Numerous cognitive components comprise successful UFOV task completion, including rapid visual attention, discrimination of targets (including visual acuity), and brief visuospatial memory (for a comprehensive review and meta-analysis, see (Woutersen et al. 2017). Assessing the underpinnings of base components underlying the UFOV test may provide unique insight into the neural correlates of UFOV performance and potential avenues for intervention optimization.

Materials and Methods

Participants

The sample of 256 healthy older adults were recruited at the University of Florida (Gainesville, FL, USA) and University of Arizona (Tucson, AZ, USA) as part of an ongoing, randomized controlled clinical trial (R01AG054077). Participants were excluded from the study if they: were outside the age range of 65–89, had a history of head injury with loss of consciousness greater than 20 min, had a history of neurological or major psychiatric disorders, had any major, uncontrolled medical condition, were left-handed, or had MRI contraindications. The full list of inclusion/exclusion criteria are provided elsewhere (Woods et al. 2018). General cognitive ability was measured at screening using the Uniform Data Set, a widely validated tool created by the National Alzheimer’s Coordinating Center to assess MCI and Alzheimer’s disease (Weintraub et al. 2009). Possible MCI was defined as age-, education- and sex-adjusted scores 1.5 standard deviations below the mean in 1 or more of the 5 following cognitive domains: general cognition, visuospatial functioning, memory, executive functioning/working memory, and language. Participants were subsequently excluded if they met the above criteria for possible MCI. Additionally, participants were further excluded from analyses if they had abnormal findings on their MRI scan (i.e., tumors, cysts).

Of the 256 participants included in the current study, 2 participants were removed due to substantial errors in the functional imaging data (errors in the first-level implicit mask). Twenty-one participants were excluded for responding to less than 75% of trials on the UFOV fMRI task. As participants were told to respond to each trial, a response rate of less than 75% suggests noncompliance or confusion with task instructions. In total, 233 participants were included in the final analyses (Table 1). The University of Florida and University of Arizona Institutional Review Boards approved all protocol procedures. All participants provided written informed consent prior to the start of study procedures, in accordance with the Declaration of Helsinki. See Table 1 for demographic information.

Table 1.

Sample demographics

	University of Florida (n = 151)	University of Arizona (n = 82)	Combined (n = 233)
Age mean (SD)	71.49 (5.43)	71.11 (4.34)	71.36 (5.07)
Education mean (SD)	16.16 (2.57)	16.45 (1.85)	16.26 (2.35)
Sex male	62	25	87
UFOV % accuracy mean (SD)	62.23 (11.51)	62.67 (12.33)	62.39 (11.78)

SD = standard deviation.

UFOV fMRI Task

The original UFOV test is a widely validated measure of speed of processing and attention, containing 4 separate subtasks. The UFOV fMRI task is based on subtask 3 of the original UFOV test, and is specifically adapted for in-scanner administration. This task requires participants to attend to a central target, as well as a peripheral target among distractors (Edwards et al. 2006). All participants completed the UFOV fMRI task in-scanner, created with E-Prime version 2.0 (Psychology Software Tools Inc., Pittsburgh, PA, USA). Responses in-scanner were made via an MRI-compatible button box, and all participants were extensively trained on this task prior to the MRI to ensure understanding.

During the “stimulus” portion of this task, the participant is required to identify the target in the center of the screen (either a car or a truck) and also attend to the location of a target (a car) in the periphery, among distractors. During the “probe” portion of the UFOV fMRI task, the participant performs a forced choice response in which they decide if the object in the center of the screen and the location of the car in the periphery are the same as the previously observed stimulus screen. The “stimulus” screen and “probe” screens are separated by a “mask”, presented for an average of 3000 ms. Individual trials of the task are separated by a “fixation” cross, presented for an average of 5000 ms. The “stimulus” screens and “probe” screens were presented for fixed durations of 500 ms and 2500 ms, respectively (Fig. 1). Within the original UFOV test, scores are a function of the briefest display duration in which participants respond correctly on at least 75% of trials, recorded in milliseconds. Given limitations on fMRI scan acquisition, the “stimulus” portion of the task was presented for a static duration of 500 ms, as opposed to an adaptive display duration. All participants received 56 trials of the UFOV fMRI task in-scanner, and the number of distractors during the “stimulus” screen range from 0 to 47 distractors. Each participant received 8 trials of 0, 7, 15, 23, 31, 39, and 47 distractors to assess various levels of task difficulty. Within-task accuracy is defined as the number of correct responses divided by 56 trials. The total run time of the UFOV fMRI task was 10 min and 32 s.

Figure 1.

Useful Field of View task.

MRI Acquisition

T1-weighted magnetization prepared rapid gradient echo (MPRAGE) structural scans were collected for coregistration via a Siemens 3 T Prisma scanner using a 64-channel head coil (University of Florida site, n = 151) or via a Siemens 3 T Skyra scanner using a 32-channel head coil (University of Arizona site, n = 82) (Siemens, Erlangen, Germany). T1-weighted MPRAGE scans were collected with the following parameters at both sites: repetition time (TR) = 1800 ms; echo time (TE) = 2.26 ms; flip angle = 8°; field of view = 256 mm × 256 mm × 176 mm; voxel size = 1 mm³. Task-based fMRI scans were collected using multislice interleaved echo planar imaging (EPI-BOLD) using the following parameters: repetition time (TR) = 3000 ms; echo time (TE) = 30.0 ms; flip angle = 70^o; field of view = 240 mm × 240 mm × 132 mm; voxel = 3 mm³.

Functional Neuroimaging Processing

Functional images were preprocessed using Statistical Parametric Mapping (SPM12) software (Wellcome Department of Imaging Neuroscience, London, UK) within the MATLAB suite (MATLAB 2016b; Mathworks, Inc., Natick, MA, USA). Preprocessing steps include motion estimation and correction, slice timing correction, artifact detection toolbox (ART) identification of motion outliers (set at the 97th percentile), tissue segmentation, normalization to MNI space and smoothing (set at 8 mm full-width half maximum Gaussian smoothing kernel). Motion information from ART was included as a covariate in all first-level analyses as nuisance variables. First-level contrasts were performed to model the hemodynamic response within participants. This hemodynamic response was modeled onto the onsets and duration of the stimulus and probe screens of the UFOV fMRI task. Group level analyses were then performed to evaluate the BOLD response in each of the following contrasts: “Stimulus” > “probe”, “probe” > “stimulus” and the conjunction analysis of “stimulus” > “rest and probe” > “rest”. One-sample t-tests were used to assess if BOLD activation patterns within the sample differed from the mean. Age, education, sex, and scanner location were included as covariates in all second-level models. Significant clusters were defined as a contiguous cluster extent threshold of k = 10 voxels, and all clusters were corrected at a false discovery rate (FDR) threshold at P < 0.05 (Benjamini and Hochberg 1995). The total percent of regions of interest (ROI) encompassed was defined as the number of voxels with significant BOLD activation divided by the total number of voxels within each ROI, using the AAL atlas labels. All contrasts were then rerun with accuracy included as a covariate the model to assess the impact of performance on BOLD response, using the above parameters.

Results

Widespread whole-brain bilateral activation in the frontal, temporal, parietal, and occipital lobes, as well as bilateral cerebellum, central structures, insula, and cingulate gyri was evidenced in the “stimulus” > “probe” contrast. This contrast consisted of 3 significant clusters centered at the following: left inferior occipital gyrus (t(228) = 16.00, pFDR < 2.7944e−14), right cingulate gyrus (t(228) = 5.64, pFDR = 8.6975e−08), and right inferior temporal gyrus (t(228) = 2.94, pFDR = 0.0036) (Fig. 2,  Table 2).

Figure 2.

Activation locations in the “stimulus” > “probe” contrast.

Table 2.

Activation locations in the “stimulus” > “probe” contrast¹

Cluster extend	#voxels (L)	#voxels (R)
Stimulus > probe Condition (3 clusters)
Inferior occipital gyrus (L) (k = 94 871)(x = −40, y = −60, z = −8)
Frontal lobe
Middle frontal gyrus	2042 (41.99%)	2880 (56.43%)
Precentral gyrus	2584 (73.28%)	1461 (43.21%)
Supplementary motor area	1405 (65.44%)	1031 (43.48%)
Superior frontal gyrus	889 (24.70%)	1201 (29.61%)
Inferior frontal gyrus, pars triangularis	1154 (45.63%)	956 (44.44%)
Inferior frontal gyrus, pars opercularis	765 (73.70%)	1059 (75.70%)
Superior frontal gyrus, medial region	316 (10.56%)	174 (8.15%)
Paracentral lobule	149 (11.05%)	206 (24.64%)
Inferior frontal gyrus, pars orbitalis	167 (9.88%)	198 (11.60%)
Middle frontal gyrus, orbital region	167 (18.81%)	119 (11.72%)
Rolandic operculum	144 (14.55%)	78 (5.86%)
Superior frontal gyrus, orbital region	132 (13.71%)	135 (13.54%)
Superior frontal gyrus, medial orbital region	–	32 (3.72%)
Gyrus rectus	7 (0.82%)	30 (4.03%)
Olfactory cortex	3 (1.07%)	26 (9.00%)
Temporal lobe
Middle temporal gyrus	1171 (23.69%)	1600 (36.29%)
Fusiform gyrus	1472 (63.72%)	1529 (60.72%)
Superior temporal gyrus	248 (10.80%)	589 (18.75%)
Inferior temporal gyrus	360 (11.25%)	389 (10.94%)
Parahippocampal gyrus	301 (30.78%)	382 (33.75%)
Hippocampus	368 (39.48%)	373 (39.43%)
Temporal pole	82 (6.38%)	24 (1.79%)
Amygdala	19 (8.64%)	48 (19.35%)
Heschl’s gyrus	4 (1.78%)	–
Parietal lobe
Postcentral gyrus	2287 (58.76%)	1660 (43.42%)
Precuneus	1722 (48.81%)	2175 (66.62%)
Inferior parietal lobule	1797 (73.44%)	840 (62.45%)
Superior parietal lobule	953 (46.15%)	779 (35.06%)
Supramarginal gyrus	251 (19.98%)	681 (34.50%)
Angular gyrus	174 (14.83%)	544 (31.05%)
Occipital lobe
Middle occipital gyrus	1929 (58.99%)	859 (40.94%)
Lingual gyrus	1763 (84.15%)	1879 (81.70%)
Calcarine fissure and surrounding cortex	1468 (65.01%)	1488 (79.96%)
Cuneus	837 (54.85%)	902 (63.34%)
Superior occipital gyrus	783 (57.32%)	763 (54.00%)
Inferior occipital gyrus	402 (42.72%)	211 (21.33%)
Cerebellum
Lobule VI of the cerebellum	1602 (94.57%)	1598 (89.03%)
Lobule VI of the cerebellum	1602 (94.57%)	1598 (89.03%)
Crus I of the cerebellum	991 (38.07%)	601 (22.70%)
Lobules IV/V of the cerebellum	936 (83.20%)	760 (88.27%)
Vermis lobules IV/V of the cerebellum	665 (100.00%)
Vermis lobule VI of the cerebellum	365 (98.38%)
Vermis lobule III of the cerebellum	157 (68.86%)
Vermis lobule VII of the cerebellum	146 (75.26%)
Vermis lobule VIII of the cerebellum	102 (41.98%)
Vermis lobule IX of the cerebellum	100 (57.47%)
Lobule IX of the cerebellum	72 (8.29%)	74 (9.15%)
Lobule III of the cerebellum	66 (48.53%)	71 (34.30%)
Lobule X of the cerebellum	62 (43.06%)	42 (26.42%)
Crus II of the cerebellum	49 (2.59%)	38 (1.79%)
Lobule VIII of the cerebellum	46 (2.44%)	33 (1.43%)
Lobule VII of the cerebellum	29 (4.96%)	–
Vermis lobule X of the cerebellum	20 (17.86%)
Vermis lobule I/II of the cerebellum	3 (5.66%)
Central structures
Thalamus	1072 (97.45%)	1036 (98.01%)
Putamen	633 (62.74%)	617 (57.99%)
Caudate nucleus	500 (51.98%)	633 (63.68%)
Pallidum	222 (75.77%)	200 (71.43%)
Insula and cingulate gyri
Insula	699 (37.62%)	595 (33.62%)
Middle cingulate gyrus	525 (27.05%)	637 (28.92%)
Anterior cingulate gyrus	152 (10.86%)	179 (13.63%)
Posterior cingulate gyrus	45 (9.72%)	78 (23.28%)
Inferior temporal gyrus (R)(k = 29)(x = 38, y = −4, z = −26)
Temporal lobe
Inferior temporal gyrus	–	2 (0.06%)
Hippocampus	–	2 (0.21%)
Amygdala	–	2 (0.81%)
Fusiform gyrus	–	1 (0.04%)
Cingulate gyrus (R)(k = 275)(x = 4, y = −32, z = 28)
Insula and cingulate gyri
Middle cingulate gyrus	23 (1.18%)	99 (4.49%)
Posterior cingulate gyrus	15 (3.24%)	9 (2.69%)

¹Corrected at voxelwise FDR P < 0.05 cluster locations are italicized, MNI coordinates are listed for peak activation at each cluster and k denotes the number of significant voxels within that cluster. Activation of other tissue types (i.e., white matter) are not included in the table, but are listed in total number of significant voxels (k). Percentage label defined as the number of significant voxels within an ROI divided by the total number of voxels in that ROI.

Conversely, the “probe” > “stimulus” contrast was associated with discrete activation patterns within primarily regions of the frontal, temporal, and parietal lobes, as well as the insula, and the cingulate cortex. This contrast consisted of 13 significant clusters, centered at the following: left inferior temporal gyrus (t(228) = 4.02, pFDR = 0.0026), left parahippocampal gyrus (t(228) = 3.54, pFDR = 0.0096), right middle temporal gyrus (t(228) = 3.52, pFDR = 0.01), left superior frontal gyrus-medial region (t(228) = 7.16, pFDR = 2.0147e−07), left and right pars orbitalis (t(228) = 5.07, pFDR = 9.3304e−05 and t(228) = 3.19, pFDR = 0.0235, respectively), right insula (t(228) = 5.55, pFDR = 1.6866e−05), left Rolandic operculum (t(228) = 7.30, pFDR = 2.0147e−07), left posterior cingulate gyrus (t(228) = 4.96, pFDR = 1.4115e−04), left and right angular gyrus (t(228) = 6.94, pFDR = 2.0147e−07 and t(228) = 4.00, pFDR = 0.0028, respectively), right superior frontal gyrus (t(228) = 4.12, pFDR = 0.002), and right postcentral gyrus (t(228) = 3.35, pFDR = 0.0156) (Fig. 3,  Table 3).

Figure 3.

Activation locations in the “probe” > “stimulus” contrast.

Table 3.

Activation locations in the “probe” > “stimulus” contrast¹

Cluster extend	#voxels (L)	#voxels (R)
Probe > stimulus condition (13 clusters)
Inferior temporal gyrus (L) (k = 147)(x = −52, y = −18, z = −26)
Temporal lobe
Middle temporal gyrus	75 (1.52%)	–
Inferior temporal gyrus	72 (2.25%)	–
Parahippocampal gyrus (L) (k = 14)(x = −26, y = −10, z = −26)
Temporal lobe
Hippocampus	10 (1.07%)	–
Parahippocampal gyrus	4 (0.41%)	–
Middle temporal gyrus (R) (k = 11)(x = 56, y = −8, z = −24)
Temporal lobe
Middle temporal gyrus	–	11 (0.25%)
Superior frontal gyrus, medial region (L)(k = 3138)(x = −8, y = 52, z = 16)
Frontal lobe
Superior frontal gyrus, medial region	854 (28.54%)	486 (22.77%)
Superior frontal gyrus	394 (10.95%)	56 (1.38%)
Superior frontal gyrus, medial orbital	72 (10.01%)	3 (0.35%)
Gyrus rectus	62 (7.28%)	4 (0.54%)
Olfactory cortex	23 (8.21%)	1 (0.35%)
Middle frontal gyrus	5 (0.10%)	–
Insula and cingulate gyri
Anterior cingulate gyrus	640 (45.71%)	378 (28.79%)
Middle cingulate gyrus	–	4 (0.18%)
Inferior frontal gyrus, pars orbitalis (L)(k = 169)(x = −34, y = 32, z = −14)
Frontal lobe
Inferior frontal gyrus, pars orbitalis	154 (9.11%)	–
Inferior frontal gyrus, pars triangularis	13 (0.51%)	–
Middle frontal gyrus, orbital region	2 (0.23%)	–
Inferior frontal gyrus, pars orbitalis (R) (k = 10)(x = 46, y = 36, z = −12)
Frontal lobe
Inferior frontal gyrus, pars orbitalis	–	10 (0.59%)
Insula (R)(k = 980)(x = 38, y = −16, z = 6)
Frontal lobe
Rolandic operculum	–	363 (27.27%)
Heschl’s gyrus	–	88 (35.34%)
Temporal lobe
Superior temporal gyrus	–	104 (3.31%)
Parietal lobe
Supramarginal gyrus	–	36 (1.82%)
Insula and cingulate gyri
Insula	–	321 (18.14%)
Rolandic operculum (L)(k = 780)(x = −38, y = −18, z = 16)
Frontal Lobe
Rolandic operculum	245 (24.75%)	–
Heschl’s gyrus	54 (24.00%)	–
Temporal lobe
Superior temporal gyrus	155 (6.75%)	–
Parietal lobe
Supramarginal gyrus	7 (0.56%)	–
Insula and cingulate gyri
Insula	300 (16.15%)	-
Posterior cingulate gyrus (L) (k = 781)(x = −6, y = −46, z = 32)
Parietal lobe
Precuneus	231 (6.55%)	48 (1.47%)
Insula and cingulate gyri
Posterior cingulate gyrus	229 (49.46%)	82 (24.48%)
Middle cingulate gyrus	108 (5.56%)	59 (2.68%)
Angular gyrus (L) (k = 344)(x = −52, y = −64, z = 30)
Temporal lobe
Middle temporal gyrus	33 (0.67%)	–
Parietal lobe
Angular gyrus	310 (26.43%)	–
Inferior parietal lobule	1 (0.04%)	–
Angular gyrus (R) (k = 26)(x = 54, y = −58, z = 32)
Parietal lobe
Angular gyrus	–	25 (1.43%)
Inferior parietal lobule	–	1 (0.07%)
Superior frontal gyrus (R)(k = 65)(x = 16, y = 38, z = 48)
Frontal lobe
Superior frontal gyrus	–	54 (1.33%)
Superior frontal gyrus, medial region	–	11 (0.52%)
Postcentral gyrus (R)(k = 16)(x = 18, y = −34, z = 68)
Parietal lobe
Postcentral gyrus	–	15 (0.39%)
Frontal lobe
Precentral gyrus	–	1 (0.03%)

¹Corrected at voxelwise FDR P < 0.05 cluster locations are italicized, MNI coordinates are listed for peak activation at each cluster and k denotes the number of significant voxels within that cluster. Activation of other tissue types (i.e., white matter) are not included in the table, but are listed in total number of significant voxels (k). Percentage label defined as the number of significant voxels within an ROI divided by the total number of voxels in that ROI.

When assessing the “stimulus” > “rest and probe” > “rest conjunction analysis”, again we observed bilateral, near whole-brain shared activation patterns within the frontal, parietal, temporal, and occipital lobes, as well as the cerebellum, insula, and thalamus. This contrast consisted of 7 significant clusters, centered at the following: left middle occipital gyrus (t(460) = 13.92, pFDR < 2.7944e−14), left precentral gyrus (t(460) = 11.86, pFDR < 2.7944e−14), right insula (t(460) = 9.53, pFDR < 2.7944e−14), right IPL (t(460) = 7.71, pFDR = 3.7690e−10), left and right thalamus (t(460) = 5.61, pFDR = 4.2261e−05 and t(460) = 6.46, pFDR = 5.6532e−07, respectively), and left middle frontal gyrus (t(460) = 4.97, pFDR = 6.8826e−04) (Fig. 4, Table 4).

Figure 4.

Activation locations in the “stimulus” > “rest and probe” > “rest” conjunction analysis.

Table 4.

Activation locations in the “stimulus” > “rest and probe” > “rest” conjunction analysis¹

Cluster extend	#voxels (L)	#voxels (R)
Stimulus + probe > rest conjunction (8 clusters)
Middle occipital gyrus (L) (k = 10 815)(x = −20, y = 94, z = 0)
Temporal lobe
Fusiform gyrus	855 (37.01%)	1218 (48.37%)
Inferior temporal gyrus	57 (1.78%)	113 (3.18%)
Parahippocampal gyrus	–	15 (1.33%)
Middle temporal gyrus	–	4 (0.09%)
Occipital lobe
Lingual gyrus	889 (42.43%)	889 (38.65%)
Calcarine fissure and surrounding cortex	624 (27.64%)	546 (29.34%)
Middle occipital gyrus	415 (12.69%)	153 (7.29%)
Inferior occipital gyrus	309 (32.84%)	155 (15.67%)
Superior occipital gyrus	139 (10.18%)	96 (6.79%)
Cuneus	90 (5.90%)	122 (8.57%)
Cerebellum
Lobule VI of the cerebellum	1026 (60.57%)	1296 (72.20%)
Crus I of the cerebellum	443 (17.02%)	171 (6.46%)
Lobules IV/V of the cerebellum	–	333 (38.68%)
Vermis lobule VI of the cerebellum	140 (37.74%)
Vermis lobule IV/V of the cerebellum	22 (3.31%)
Vermis lobule VII of the cerebellum	17 (8.76%)
Lobule IV/V of the cerebellum	16 (1.42%)	–
Crus II of the cerebellum	10 (0.53%)	3 (0.14%)
Lobule VIII of the cerebellum	1 (0.05%)	7 (0.30%)
Precentral gyrus (L) (k = 12 683)(x = −34, y = −22, z = 56)
Frontal lobe
Precentral gyrus	2045 (58.00%)	–
Supplementary motor area	974 (45.37%)	442 (18.64%)
Inferior frontal gyrus, pars triangularis	898 (35.51%)	–
Inferior frontal gyrus, pars opercularis	719 (69.27%)	–
Middle frontal gyrus	616 (12.67%)	–
Superior frontal gyrus, medial region	311 (10.39%)	170 (7.97%)
Superior frontal gyrus	181 (5.03%)	2 (0.05%)
Rolandic operculum	85 (8.59%)	–
Inferior frontal gyrus, pars orbitalis	73 (4.32%)	–
Paracentral lobule	31 (2.30%)	–
Parietal lobe
Inferior parietal lobule	1647 (67.31%)	–
Postcentral gyrus	1359 (34.92%)	–
Superior parietal lobule	524 (25.38%)	–
Angular gyrus	146 (12.45%)	–
Supramarginal gyrus	132 (10.51%)	–
Precuneus	126 (3.57%)	–
Occipital lobe
Middle occipital gyrus	163 (4.98%)	–
Superior occipital gyrus	4 (0.29%)	–
Insula and cingulate gyri
Insula	514 (27.66%)	–
Middle cingulate gyrus	185 (9.53%)	279 (12.66%)
Anterior cingulate gyrus	–	19 (1.45%)
Insula (R)(k = 4868)(x = 32, y = 24, z = −2)
Frontal lobe
Middle frontal gyrus	–	1499 (29.37%)
Inferior frontal gyrus, pars opercularis	–	883 (63.12%)
Inferior frontal gyrus, pars triangularis	–	782 (36.36%)
Precentral gyrus	–	589 (17.42%)
Superior frontal gyrus	–	181 (4.46%)
Inferior frontal gyrus, pars orbitalis	–	105 (6.15%)
Rolandic operculum	–	7 (0.53%)
Middle frontal gyrus, orbital region	–	4 (0.39%)
Superior frontal gyrus, orbital region	–	1 (0.10%)
Insula and cingulate gyri
Insula	–	379 (21.41%)
Central structures
Putamen	–	28 (2.63%)
Inferior parietal lobule (R) (k = 3319)(x = 36, y = −54, z = 46)
Frontal lobe
Precentral gyrus	–	1 (0.03%)
Parietal lobe
Inferior parietal lobule	–	770 (57.25%)
Angular gyrus	–	498 (28.42%)
Postcentral gyrus	–	481 (12.58%)
Supramarginal gyrus	–	353 (17.88%)
Superior parietal lobule	–	281 (12.65%)
Precuneus	–	173 (5.30%)
Occipital lobe
Middle occipital gyrus		156 (7.44%)
Superior occipital gyrus		57 (4.03%)
Thalamus (L) (k = 381)(x = −22, y = −24, z = 2)
Temporal lobe
Hippocampus	2 (0.21%)	–
Central structures
Thalamus	249 (22.64%)	–
Middle frontal gyrus (L) (k = 345)(x = −40, y = 5-, z = −2)
Frontal lobe
Middle frontal gyrus	178 (3.66%)	–
Middle frontal gyrus, orbital region	105 (11.82%)	–
Inferior frontal gyrus, pars triangularis	17 (0.67%)	–
Superior frontal gyrus	16 (0.44%)	–
Superior frontal gyrus, orbital region	10 (1.04%)	–
Inferior frontal gyrus, pars orbitalis	5 (0.30%)	–
Thalamus (R)(k = 239)(x = 24, y = −26, z = 0)
Temporal lobe
Hippocampus	10 (1.06%)	–
Central structures
Thalamus	49 (4.64%)	–
Middle Frontal Gyrus, Orbital Region (L)(k = 12)(x = −20, y = 44, z = −16)
Frontal lobe
Superior frontal gyrus, orbital region	7 (0.73%)	–
Middle frontal gyrus, orbital region	5 (0.56%)	–

¹Corrected at voxelwise FDR P < 0.05 cluster locations are italicized, MNI coordinates are listed for peak activation at each cluster and k denotes the number of significant voxels within that cluster. Activation of other tissue types (i.e., white matter) are not included in the table, but are listed in total number of significant voxels (k). Percentage label defined as the number of significant voxels within an ROI divided by the total number of voxels in that ROI.

After assessing the BOLD activation patterns for each of the above contrasts, we reran the models with accuracy included to assess how UFOV fMRI task performance associated with activation patterns in each of the above contrasts. Specifically, we found 6 significant clusters, centered at: right superior parietal lobule (t(227) = 6.21, pFDR = 4.0900e−05), left and right superior frontal gyrus-orbital region (t(227) = 2.74, pFDR = 0.0205 and t(227) = 2.98, pFDR = 0.0128), left middle frontal gyrus (t(227) = 3.10, pFDR = 0.0100), left insula (t(227) = 2.75, pFDR = 0.0203), and right middle frontal gyrus (t(227) = 2.59, pFDR = 0.0272) (Fig. 5,  Table 5) Surprisingly, no clusters survived FDR correction in the “probe” > “stimulus with accuracy” contrast. This suggests that once accuracy is taken into account, there are no significant BOLD clusters that are greater in the probe condition over the stimulus condition at pFDR = 0.05.

Figure 5.

Activation locations in the “stimulus” > “probe with accuracy” contrast.

Table 5.

Activation locations in the “stimulus” > “probe with accuracy” contrast¹

Cluster extend	#voxels (L)	#voxels (R)
Stimulus > probe condition with accuracy (6 clusters)
Superior parietal lobule (R)(k = 38 872)(x = 28, y = −64, z = 52)
Frontal lobe
Precentral gyrus	1886 (53.49%)	614 (18.16%)
Middle frontal gyrus	970 (19.95%)	1337 (26.20%)
Supplementary motor area	1105 (51.47%)	363 (15.31%)
Inferior frontal gyrus, pars triangularis	739 (29.22%)	568 (26.41%)
Inferior frontal gyrus, pars opercularis	509 (49.04%)	605 (43.25%)
Superior frontal gyrus	605 (16.81%)	456 (11.24%)
Superior frontal gyrus, medial region	262 (8.76%)	123 (5.76%)
Rolandic operculum	64 (6.46%)	52 (3.91%)
Inferior frontal gyrus, pars orbitalis	3 (0.18%)	15 (0.88%)
Middle frontal gyrus, orbital region	–	3 (0.30%)
Paracentral lobule	–	1 (0.12%)
Temporal lobe
Fusiform gyrus	858 (37.14%)	606 (24.07%)
Inferior temporal gyrus	280 (8.75%)	365 (10.26%)
Middle temporal gyrus	327 (6.62%)	364 (8.26%)
Parahippocampal gyrus	11 (1.12%)	106 (9.36%)
Superior temporal gyrus	38 (1.66%)	62 (1.97%)
Hippocampus	–	8 (0.85%)
Parietal lobe
Precuneus	827 (23.44%)	1476 (45.21%)
Inferior parietal lobule	1248 (51.00%)	620 (46.10%)
Postcentral gyrus	1000 (25.69%)	401 (10.49%)
Superior parietal lobule	856 (41.45%)	557 (25.07%)
Angular gyrus	21 (1.79%)	277 (15.81%)
Supramarginal gyrus	25 (1.99%)	124 (6.28%)
Occipital lobe
Middle occipital gyrus	1224 (37.43%)	502 (23.93%)
Lingual gyrus	586 (27.97%)	865 (37.61%)
Calcarine fissure and surrounding cortex	609 (26.97%)	728 (39.12%)
Superior occipital gyrus	429 (31.41%)	416 (29.44%)
Cuneus	157 (10.29%)	308 (21.63%)
Inferior occipital gyrus	173 (18.38%)	126 (12.74%)
Cerebellum
Lobule VI of the cerebellum	1436 (84.77%)	1215 (67.69%)
Crus I of the cerebellum	549 (21.09%)	169 (6.38%)
Lobules IV/V of the cerebellum	344 (30.58%)	299 (34.73%)
Vermis lobule VI of the cerebellum	275 (74.12%)
Vermis lobule VII of the cerebellum	120 (61.86%)
Vermis lobule IV/V of the cerebellum	115 (17.29%)
Vermis lobule VIII of the cerebellum	96 (39.51%)
Vermis lobule IX of the cerebellum	90 (51.72%)
Lobule IX of the cerebellum	40 (4.60%)	33 (4.08%)
Lobule X of the cerebellum	22 (15.28%)	38 (23.90%)
Lobule VIII of the cerebellum	34 (1.80%)	32 (1.39%)
Crus II of the cerebellum	13 (0.69%)	16 (0.76%)
Lobule VII of the cerebellum	11 (1.88%)	–
Lobule III of the cerebellum	–	4 (1.93%)
Central structures
Thalamus	145 (13.18%)	172 (16.27%)
Putamen	20 (1.98%)	59 (5.55%)
Caudate nucleus	55 (5.72%)	13 (1.31%)
Pallidum	11 (3.75%)	3 (1.07%)
Insula and cingulate gyri
Insula	159 (8.56%)	141 (7.97%)
Middle cingulate gyrus	182 (9.38%)	212 (9.62%)
Anterior cingulate gyrus	50 (3.57%)	69 (5.26%)
Superior frontal gyrus, orbital region (L)(k = 15)(x = −18, y = 40, z = −14)
Frontal lobe
Superior frontal gyrus, orbital region	11 (1.14%)	–
Middle frontal gyrus, orbital region	1 (0.11%)	–
Superior frontal gyrus, orbital region (R)(k = 26)(x = 20, y = 46, z = −14)
Frontal lobe
Superior frontal gyrus, orbital region	–	17 (1.71%)
Middle frontal gyrus, orbital region	–	1 (0.10%)
Middle frontal gyrus (L)(k = 42)(x = −38, y = 48, z = 6)
Frontal Lobe
Middle frontal gyrus	40 (0.82%)	–
Inferior frontal gyrus, pars triangularis	2 (0.08%)	–
Insula (L)(k = 15)(x = −42, y = 16, z = 0)
Insula and cingulate gyri
Insula	13 (0.70%)	–
Frontal lobe
Inferior frontal gyrus, pars triangularis	2 (0.08%)	–
Middle frontal gyrus (R)(k = 11)(x = 28, y = 14, z = 46)
Frontal lobe
Middle frontal gyrus	–	11 (0.22%)

¹Corrected at voxelwise FDR P < 0.05 cluster locations are italicized, MNI coordinates are listed for peak activation at each cluster and k denotes the number of significant voxels within that cluster. Activation of other tissue types (i.e., white matter) are not included in the table, but are listed in total number of significant voxels (k). Percentage label defined as the number of significant voxels within an ROI divided by the total number of voxels in that ROI.

Analyzing the “stimulus” > “rest and probe” > “rest with accuracy conjunction” contrast, we see activation patterns more focally centered in the frontal and parietal lobes, as well throughout the cerebellum. This contrast consisted of 15 significant clusters, centered at the following: left precentral gyrus (t(459) = 6.28, pFDR = 3.8503e−10), right superior frontal gyrus-medial region (t(459) = 5.51, pFDR = 3.0419e−08), left and right lobule VI of the cerebellum (t(459) = 3.35, pFDR = 4.3958e−04 and t(459) = 5.28, pFDR = 9.9047e−08, respectively), left insula (t(459) = 4.80, pFDR = 7.9395e−07), right inferior gyrus, pars triangularis (t(459) = 4.78, pFDR = 1.1658e−06), left thalamus (t(459) = 4.00, pFDR = 3.6438e−05), right IPL (t(459) = 3.66, pFDR = 1.4071e−04), right postcentral gyrus, 2 clusters (t(459) = 3.60, pFDR = 1.7385e−04 and t(459) = 2.97, pFDR = 0.0016), left inferior frontal gyrus, pars opercularis (t(459) = 3.53, pFDR = 2.2564e−04), right middle frontal gyrus, orbital region (t(459) = 3.45, pFDR = 3.0865e−04), left caudate nucleus (t(459) = 3.34, pFDR = 4.5971e−04), left paracentral lobule (t(459) = 3.33, pFDR = 4.6704e−04), and middle temporal gyrus (t(459) = 3.25, pFDR = 6.1330e−04) (Fig. 6,  Table 6).

Figure 6.

Activation locations in the “stimulus” > “rest and probe” > “rest with accuracy” conjunction analysis.

Table 6.

Activation locations in the “stimulus” > “rest and probe” > “rest with accuracy” conjunction analysis¹

Cluster extend	#voxels (L)	#voxels (R)
Probe + stimulus > rest condition with accuracy (15 clusters)
Precentral gyrus (L)(k = 3550)(x = −32, y = −22, z = 70)
Frontal lobe
Precentral gyrus	866 (24.56%)	176 (5.21%)
Supplementary motor area	254 (11.83%)	99 (4.18%)
Superior frontal gyrus	67 (1.86%)	128 (3.16%)
Paracentral lobule	79 (5.86%)	–
Rolandic operculum	6 (0.61%)	–
Parietal lobe
Postcentral gyrus	1129 (29.01%)	1 (0.03%)
Inferior parietal lobule	129 (5.27%)	–
Superior parietal lobule	35 (1.69%)	–
Supramarginal gyrus	24 (1.91%)	–
Insula and cingulate gyri
Middle cingulate gyrus	125 (6.44%)	–
Superior frontal gyrus, medial region (R)(k = 1185)(x = 6, y = 32, z = 42)
Frontal lobe
Supplementary motor area	258 (12.02%)	318 (13.41%)
Superior frontal gyrus, medial region	244 (8.16%)	227 (10.64%)
Superior frontal gyrus	2 (0.06%)	26 (0.64%)
Insula and cingulate gyri
Middle cingulate gyrus	–	98 (4.45%)
Anterior cingulate gyrus	1 (0.05%)	3 (0.23%)
Lobule VI of the cerebellum (R)(k = 4196)(x = −30, y = −72, z = −22)
Temporal lobe
Fusiform gyrus	135 (5.84%)	385 (15.29%)
Inferior temporal gyrus	–	1 (0.03%)
Occipital lobe
Lingual gyrus	237 (11.31%)	564 (24.52%)
Calcarine fissure and surrounding cortex	88 (3.90%)	237 (12.74%)
Middle occipital gyrus	154 (4.71%)	81 (3.86%)
Inferior occipital gyrus	105 (11.16%)	68 (6.88%)
Superior occipital gyrus	18 (1.32%)	27 (1.91%)
Cuneus	–	26 (1.83%)
Cerebellum
Lobule VI of the cerebellum	452 (26.68%)	716 (39.89%)
Crus I of the cerebellum	295 (11.33%)	58 (2.19%)
Lobules IV/V of the cerebellum	–	162 (18.82%)
Vermis lobule VI of the cerebellum	95 (25.61%)
Vermis lobule IV/V of the cerebellum	20 (3.01%)
Vermis lobule VII of the cerebellum	4 (2.06%)
Lobule VIII of the cerebellum	3 (0.16%)	–
Crus II of the cerebellum	2 (0.11%)	–
Insula (L)(k = 257)(x = −30, y = 22, z = −4)
Frontal lobe
Inferior frontal gyrus, pars orbitalis	26 (1.54%)	–
Inferior frontal gyrus, pars triangularis	1 (0.04%)	–
Insula and cingulate gyri
Insula	196 (10.55%)	–
Inferior frontal gyrus, pars triangularis (R)(k = 1247)(x = 44, y = 22, z = 26)
Frontal lobe
Middle frontal gyrus	–	611 (11.97%)
Inferior frontal gyrus, pars triangularis	–	282 (13.11%)
Inferior frontal gyrus, pars opercularis	–	226 (16.15%)
Precentral gyrus	–	67 (1.98%)
Thalamus (L)(k = 154)(x = −22, y = −26, z = 0)
Central structures
Thalamus	–	111 (10.09%)
Inferior parietal lobule (R)(k = 442)(x = 42, y = −42, z = 48)
Parietal lobe
Inferior parietal lobule	–	308 (22.90%)
Angular gyrus	–	81 (4.62%)
Supramarginal gyrus	–	32 (1.62%)
Superior parietal lobule	–	2 (0.09%)
Postcentral gyrus (R)(k = 104)(x = 32, y = −28, z = 48)
Frontal lobe
Precentral gyrus	–	4 (0.12%)
Parietal lobe
Postcentral gyrus	–	87 (2.28%)
Inferior frontal gyrus, pars opercularis (L)(k = 494)(x = −48, y = 12, z = 30)
Frontal lobe
Inferior frontal gyrus, pars triangularis	183 (7.24%)	–
Middle frontal gyrus	172 (3.54%)	–
Inferior frontal gyrus, pars opercularis	83 (8.00%)	–
Precentral gyrus	56 (1.59%)	–
Middle frontal gyrus, orbital region (R)(k = 34)(x = 28, y = 52, z = −4)
Frontal lobe
Middle frontal gyrus, orbital region	–	17 (1.67%)
Superior frontal gyrus, orbital region	–	10 (1.00%)
Middle frontal gyrus	–	7 (0.14%)
Lobule VI of the cerebellum (L)(k = 78)(x = −28, y = −46, z = −22)
Temporal lobe
Fusiform gyrus	50 (2.16%)	–
Inferior temporal gyrus	1 (0.03%)	–
Cerebellum
Lobules IV/V of the cerebellum	14 (1.24%)	–
Lobule VI of the cerebellum	13 (0.77%)	–
Caudate nucleus (L)(k = 38)(x = −10, y = 10, z = −6)
Central structures
Pallidum	25 (8.53%)	–
Putamen	4 (0.40%)	–
Caudate nucleus	4 (0.42%)	–
Paracentral lobule (L)(k = 48)(x = −6, y = −36, z = 68)
Frontal lobe
Paracentral lobule	27 (2.00%)	–
Parietal lobe
Precuneus	21 (0.60%)	–
Middle temporal gyrus (R)(k = 15)(x = 62, y = −30, z = −10)
Temporal lobe
Middle temporal gyrus	–	15 (0.34%)
Postcentral gyrus (R)(k = 20)(x = 42, y = −24, z = 34)
Parietal lobe
Postcentral gyrus	–	19 (0.50%)

¹Corrected at voxelwise FDR P < 0.05 cluster locations are italicized, MNI coordinates are listed for peak activation at each cluster and k denotes the number of significant voxels within that cluster. Activation of other tissue types (i.e., white matter) are not included in the table, but are listed in total number of significant voxels (k). Percentage label defined as the number of significant voxels within an ROI divided by the total number of voxels in that ROI.

Discussion

Accumulating evidence has demonstrated durable gains from UFOV interventions in near and far transfer effects as well as improvements in cognition, lasting up to 10 years (Edwards et al. 2017). This is perhaps unsurprising, given that successful UFOV performance entails integration of a wide range of cognitive abilities, including rapid visual perception, attention, encoding, and recognition of visual targets. Despite its efficacy, few studies have assessed the neural underpinnings of this task. The current study aimed to elucidate the neural mechanisms of a UFOV-based fMRI task in a cohort of healthy older adults. Given the paucity of research into the different components of the original UFOV test, we opted for a whole-brain approach, rather than constricting our analyses to possibly irrelevant ROIs. The full extent of activation, including the percentages of the total ROI activated, can be found in Tables 2–6.

Parallels to Prior Structural Findings of the UFOV Task

Within the “stimulus” > “probe” contrast, we observed activation throughout the bilateral precentral gyrus (left = 73.28%, right = 43.21%; Table 2). The precentral gyrus contains the motor cortex and is canonically activated during motor tasks. However, research suggests that the precentral gyrus is also heavily involved in motor planning (Svoboda and Li 2018). Notably, previous research assessing the structural neural correlates of the Double Decision/UFOV task found that decreased cortical thickness in the precentral gyrus was associated with poorer performance, suggesting the precentral gyrus is essential in task completion (Kraft et al. 2020). Activations of these regions may be associated with planned motor response during the stimulus screen and the rapid presentation of the stimulus, compared with the probe screen.

Additionally, the present study found activation of the superior (left = 10.80%, right = 18.75%) and inferior (left = 11.25%, right = 10.94%) temporal gyrus within the “stimulus” > “probe” contrast (Table 2). Prior research has suggested that the superior temporal gyrus is involved with numerous cognitive domains, including attention and speed of processing (Achiron et al. 2013). These functional findings coincide with previous structural findings from Kraft et al. (2020) who demonstrated that significant cortical thinning of the fusiform gyrus, superior temporal gyrus, and inferior temporal gyrus were all associated with poorer UFOV/Double Decision performance. Combined, these findings cumulatively suggest that structural integrity of these regions may be implicated in UFOV completion.

Within the temporal lobe, considerable activation of the bilateral fusiform gyrus (left = 63.72%, right = 60.72%) was associated with the “stimulus” > “probe” condition of the UFOV fMRI task (Table 2). The fusiform gyrus is involved in highly specialized aspects of visual perception, such as object recognition and processing resolution information (Weiner and Zilles 2016; Sokolov et al. 2018). Multimodal imaging research by Dosenbach et al. (2007) has demonstrated that the fusiform gyrus, along with several different regions, including the IPL, precuneus, frontal cortex were highly functionally connected, comprising the fronto-parietal control network (FPCN). The FPCN is thought to be critically involved in UFOV performance, as greater cortical thickness in the FPCN, including the fusiform gyrus, was directly associated to better performance on the UFOV test (Schmidt et al. 2016).

Comparisons to Prior UFOV Functional Findings

We observed unique activation patterns in the bilateral SMA, located in the inferior medial region of the superior frontal gyrus. This region demonstrated considerable activation within the “stimulus” > “probe” contrast, with 65.44% of the left SMA and 43.48% of the right SMA activated (Table 2). Furthermore, this region was heavily implicated in the “stimulus” > “rest and probe” > “rest conjunction” analysis, with 45.37% of the left and 63.12% of the right SMA activated in the conjunction analyses (Table 4). When evaluating activation associated with accuracy, we observed similar levels of activation in the left SMA in the “stimulus” > “probe with accuracy” contrast (left = 51.47%, right = 15.31%; Table 5). This suggests that the SMA is essential for UFOV fMRI task completion and performance. Previous research regarding the role of the SMA lends critical insight to its association with UFOV test completion. fMRI paradigms have demonstrated that this area is not only activated during motor tasks, but is heavily involved in anticipation of completing said motor tasks and reaction time maintenance (Berman et al. 2012; Sharma and Baron 2014; Bazán et al. 2015). Additionally, Ross et al. (2018) specifically found that SMA activation was significantly decreased following a UFOV training intervention, suggesting greater neural efficiency within the SMA.

The bilateral middle frontal gyrus demonstrated unique cortical activation in the “stimulus” > “probe” contrast (left = 41.99%, right = 56.43%), as well as the orbital region of the middle frontal gyrus (left = 18.81%, right = 11.72%; Table 2). The middle frontal gyrus comprises the DLPFC and both the DLPFC and middle frontal gyrus have been linked to processing speed, particularly within functional and resting state connectivity analyses (Ivanoff et al. 2008; Rypma and Prabhakaran 2009; Ross et al. 2018; Haupt et al. 2019). Similar to the SMA, following a UFOV training intervention, activation in the DLPFC was substantially reduced, suggesting neural efficiency in this ROI (Ross et al. 2018).

Nearly the entire bilateral thalamus (left = 97.45%, right = 98.01%) was activated during the “stimulus” > “probe” contrast (Table 2). Ross et al. (2018) have previously shown significant thalamic activation in response to UFOV trials overall. These findings further delineate that thalamic activation is unique to the stimulus portion of the UFOV fMRI task. Sections of the thalamus, including the pulvinar, are responsible for visual attentional filtering, resulting in the capacity to concentrate on target items within a field of distractors, which is essential in the UFOV test (Grieve et al. 2000; Fama and Sullivan 2015).

Activation of the bilateral ACC was greater in the “probe” > “stimulus” contrast (left = 45.71%, right = 28.79%), than in the “stimulus” > “probe” contrast (left = 10.86%, right = 13.63%; Tables 3 and 2, respectively). This may suggest that the ACC is related to recall component of the UFOV fMRI task, rather than the visual perceptual components. Recent research has demonstrated that functional connectivity within the CON network, including the ACC, is essential for general processing speed performance (Haupt et al. 2019; Ruiz-Rizzo et al. 2019). Evidence has further suggested that the ACC is involved with decision making (Botvinick 2007). These findings coincide with findings by Ross et al. (2018), who found significant activation in the ACC. Moreover, BOLD activation in the ACC was attenuated after UFOV cognitive training, suggesting increased neural efficiency in this region.

Similarities with Prior Electrophysiological Findings Associated with the UFOV Test and Intervention Effects

Substantial work by O’Brien et al. (2013, 2015) have assessed electrophysiological associations with the UFOV test, specifically within P1, N2pc, and P3b, which canonically gauge early perceptual processing, visual attention, and categorization of stimuli, respectively. P1 amplitude and latency was significantly associated with performance on subtask 1 of the UFOV test, representing rapid perceptual attention and selection of a central target (O’Brien et al. 2015). The PO3 and PO4 electrode locations used in the study of P1 amplitude by O’Brien et al. align roughly with the left and right superior occipital gyri. These findings align nicely with the current study. Specifically, in the “stimulus” > “probe” contrast, when participants were actively encoding the stimulus, we observed significant activation of the left (57.32% of ROI) and right (54.00% of ROI) superior occipital gyrus. The same pattern of findings held in the current study, even after accounting for accuracy (“stimulus” > “probe with accuracy” contrast), suggesting that the superior occipital gyri are essential in UFOV fMRI test completion, likely due to its role in rapid perceptual attention.

Additionally, speed of processing training utilizing many of the same cognitive components as the UFOV test has been shown to impact N2pc and P3b amplitude. Specifically, processing speed intervention training was associated with increased amplitude of P3b, measured at Pz, contains both the superior and IPLs. Additionally, increased N2pc amplitude, measured at the same location as the above P1 findings, was associated with a processing speed intervention. In regards to the P3b amplitude, this again aligns nicely with the current findings. In the “stimulus” > “probe” contrast, during the encoding the stimulus, we see substantial activation in the left (73.44% of ROI) and right (62.45%) IPL, as well as portions of the left (46.15% of ROI) and right (35.06%) superior parietal lobule. After taking into account the variance associated with accuracy, in the “stimulus” > “probe with accuracy” contrast, activation within large portions of both the left (51.00% of ROI) and right (46.10% of ROI) IPL and portions of the left (41.45% of ROI) and right (25.07% of ROI) superior parietal remain. This further suggests that the superior and IPLs are essential for UFOV fMRI completion, likely attributable to its part in categorization of stimuli.

Associations with Both Structural and Functional Findings of the UFOV Test

Extensive functional activation of the IPL (left = 73.44%, right = 62.45%) in the “stimulus” > “probe contrast” was evidenced in the parietal lobe (Table 2). The IPL is heavily associated with spatial orientation and visual target search. The IPL, containing the temporoparietal junction, has been implicated in the FPCN (Dosenbach et al. 2007). Similar to the fusiform gyrus, greater cortical thickness in the FPCN, including the IPL, was directly associated to better performance on the UFOV test (Schmidt et al. 2016). Conversely, degradation of cortical thickness in the IPL was directly related to poorer UFOV/Double Decision performance (Kraft et al. 2020). Considerable activation was observed in the bilateral IPL in the “stimulus” > “rest and probe” > “rest conjunction analysis”, with 67.31% of the left IPL and the 57.25% of the right IPL activated (Table 4). Similar to the SMA, this region has been demonstrated to be a crucial ROI in UFOV test completion. The IPL is an integral part of the FPCN, and strengthened network connectivity within the FPCN has been associated with in-scanner UFOV subtask 2 performance and a task resembling subtask 3, respectively (Scalf et al. 2007; Ross et al. 2018). Additionally, functional activation in the IPL was significantly reduced following a UFOV-based intervention, again suggesting greater neural efficiency within the IPL (Ross et al. 2018). Considerable activation in the IPL was conserved between in “stimulus” > “probe with accuracy” contrast (left = 49.04%, right = 43.25%), compared with the original “stimulus” > “probe” contrast (left = 73.44%, right = 62.45%), suggesting that the IPL is integral in both UFOV fMRI task completion and accuracy (Tables 5 and 2, respectively).

Assessing the “stimulus” > “probe” contrast, we see significant activation within nearly all AAL atlas regions in the occipital lobe. Specifically, we see substantial activation in the bilateral superior (left = 57.32%, right = 54.00%), middle (left = 58.99%, right=) and inferior (left = 42.72%, right = 21.33%) occipital gyri, bilateral lingual gyrus (left = 84.15%, right = 81.70%), bilateral calcarine fissure and surrounding cortex (left = 65.01%, right = 79.96%), and bilateral cuneus (left = 54.85%, right = 63.34%; Table 2). Cortical thickness degradation in the lateral occipital gyrus was associated with poorer UFOV/Double Decision performance (Kraft et al. 2020). These results support the findings by Ross et al. (2018), who found significant activation in the visual cortex, specifically centered at the calcarine fissure and surrounding cortex, associated with UFOV test completion. The occipital lobe as a whole is canonically activated in visual perceptual tasks, and is activated during tasks of object recognition (Grill-Spector et al. 2001). Given the visual complexity of the stimulus screen, as well as brief visual encoding essential for task completion, significant findings throughout the occipital lobe are perhaps expected.

Slightly over one-third of both the left (37.62%) and right (33.62%) insula was implicated in the “stimulus” > “probe” contrast (Table 2). The anterior insula has been shown to be directly activated during UFOV test completion, and alterations in insular activation have been shown following UFOV cognitive training (Ross et al. 2018). In fact, some of the largest changes in task-based activation occurred within the anterior insula and SMA following a UFOV intervention, suggesting key involvement in UFOV completion (Ross et al. 2018). Indeed, insular involvement of the UFOV test has been demonstrated by previous research, suggesting that cortical thickness of the right insula was directly related to UFOV/Double Decision performance (Kraft et al. 2020). Interestingly, an additional 34.29% of the bilateral insula was implicated in the “probe” > “stimulus” contrast (left = 16.15%, right = 16.14%; Table 3). The significant activation in both the “stimulus” > “probe and probe” > “stimulus” contrasts suggests that discrete regions of the insula may be responsible for distinct components of task completion. The anterior insula is an integral component of the cingulo-opercular network (CON), also known as the ventral attention network, which has been implicated in target detection and error processing during tasks of cognition (Fox et al. 2006; Corbetta et al. 2000). Studies have shown that increased functional connectivity of the CON network, which include many of the ROIs in the present study (anterior insula, basal ganglia, thalamus, ACC, orbital gyri, and frontal gyri) was directly associated with visual processing speed performance (Haupt et al. 2019; Ruiz-Rizzo et al. 2019).

Unique Cortical Activation Patterns

In the present study, we found unique areas of activation not elucidated by previous structural or functional analyses of the UFOV fMRI task, supporting our whole-brain approach. We observed extensive activation in the bilateral pars opercularis (left = 73.70%, right = 75.70%) in the “stimulus” > “probe contrast” (Table 2). Interestingly, we see a similar pattern of activation in the “stimulus” > “rest and probe” > “rest conjunction analysis”, with 69.27% of the left pars opercularis and 63.12% of the right pars opercularis activated in this contrast (Table 3). The conservation of activation patterns suggests that activation within the pars opercularis may be associated with both the perceptual and recall components of the UFOV fMRI task. Additionally, extensive activation was maintained in the pars opercularis in the “stimulus” > “probe with accuracy” contrast (left = 49.04%, right = 43.25%), suggesting that activation in the pars opercularis is a key component associated with UFOV fMRI performance, even after evaluating the variance associated with accuracy (Table 5). The pars opercularis is typically thought of in the context of language production, as it comprises part of Broca’s area. However, research spanning the past several decades has elucidated the role of the pars opercularis in motor imagery tasks and preparation of motor tasks, particularly goal-oriented fine motor tasks (Koski et al. 2002; Johnson-Frey et al. 2003; Molnar-Szakacs et al. 2005). As such, significant activation of the pars opercularis may be due to an anticipatory response in both the “stimulus” > “rest and probe” > “rest” contrasts, as well as in their conjunction analysis (Tables 2–4).

We observed substantial activation thorough the entire cerebellum during the “stimulus” > “probe” contrast (Table 2). These significant regions of activation include bilateral lobule IV/V of the cerebellum (left = 83.20%, right = 88.27%), bilateral lobule VI of the cerebellum (left = 94.57%, right = 89.03%). In addition, we see considerable activation throughout the vermis lobules, comprising of vermis lobule III (68.86%), IV/V (100%), VI (98.38%), VII (75.26%), VIII (41.98%), IX (57.47%), and X (17.86%) of the cerebellum. Unexpectedly, we see comparable levels of activation in lobule IV of the cerebellum in the “stimulus” > “probe with accuracy” contrast (left = 84.77%, right = 67.69%), compared with the original “stimulus” > “probe” contrast (Tables 5 and 2, respectively). This suggests that lobule IV of the cerebellum in particular, plays a considerable role in UFOV fMRI task performance, again even after evaluating the variance associated with accuracy. Although the cerebellum was previously thought to be associated solely with motor control, a growing body of literature suggests the cerebellum is in fact essential to various cognitive and affective systems. Research has shown that functional connectivity from the cerebellum to frontal and parietal regions are essential in processing speed tasks (Wong et al. 2021). Lobule IV of the cerebellum in particular has also been associated with right-handed finger tapping tasks, a key component of task completion in the current study (Stoodley et al. 2012).

Limitations and Future Directions

The present study aimed to assess BOLD activation patterns associated with the completion of a UFOV fMRI-based task and behavioral performance. Even though we were able to calculate the percentage of ROI activated in the accuracy contrasts, this does not necessarily correlate to its relative importance in task performance. For instance, although we observed 51.47% activation of the left SMA in the “stimulus” > “probe with accuracy” contrast, we cannot definitively say that the SMA contributed more to UFOV fMRI task performance than the left insula, which demonstrated 8.56% activation. Perhaps the small portion of the left insula provides more significant contributions to differences in UFOV fMRI accuracy, however, this cannot be elucidated from the current study. Nevertheless, calculation of percent activation of regions provides a relative metric of the extent of activation of brain regions associated with the different components of UFOV fMRI task.

Limitations on fMRI scan acquisition prevent the original UFOV test from being administered in-scanner. The original UFOV test performance is based on the briefest presentation time in which participants accurately respond to 75% of trials. This results in hemodynamic responses that are difficult to deconvolve in rapid event-related fMRI. Additionally, alteration of presentation time based on intraindividual responses makes calibration of scan cards between participants impossible. Nevertheless, the current study attempted to mirror subtask 3 of the UFOV test as closely as possible within a fMRI environment.

One significant limitation of the current paper is the inherent limitation of spatial resolution in BOLD imaging. We are unable to delineate activation patterns of smaller nuclei within an ROI, such as the lateral geniculate nucleus of the thalamus. Additionally, we used a unimodal neuroimaging approach to assess BOLD activation patterns associated with the UFOV test. Future research should include multimodal imaging, such as structural MRI and DTI to assess how volume and white matter integrity, combined with cortical activation influence UFOV performance.

Our sample was highly educated, with a mean education above a bachelor’s degree (16.26 years), and the vast majority self-identified as white, non-Hispanic (88.46% of our sample). As such, our results may not be entirely generalizable to other racial and ethnic groups. Additionally, it is widely known that cognitive reserve is a significant factor in the preservation of cognition in later life, which could in turn, affect our results (Stern 2013). Future research should investigate BOLD activation patterns and UFOV performance within diverse racial and ethnic backgrounds, as well as within a wider range of education.

Conclusion

To date, this is the largest study assessing BOLD activation patterns during UFOV fMRI task completion. We demonstrated unique cortical activation patterns associated with both the encoding (stimulus) and recall (probe) components of the UFOV fMRI task. Notably, we further demonstrated that accuracy was associated with differences in activation patterns of the UFOV fMRI task, particularly within the “stimulus” > “probe” and conjunction contrasts. Although our study investigated an fMRI-adapted UFOV task, it provides critical insight into neural underpinnings of performance on components central to the UFOV cognitive training paradigm. These findings may provide potential avenues for intervention optimization, further increasing the efficacy of UFOV intervention in slowing age-related cognitive decline.

Funding

National Institute on Aging (NIA R01AG054077); the National Institutes of Health (NIH T32AG061892, NIH T32HL134621, NIH T32AG020499); the State of Arizona and Arizona Department of Health Services (ADHS); and the McKnight Brain Research Foundation.