http://orcid.org/0000-0002-2926-6457
Figures
Abstract
The trail-making test (TMT) is a popular neuropsychological test, which is used extensively to measure cognitive impairment associated with neurodegenerative disorders in older adults. Behavioural performance on the TMT has been investigated in older populations, but there is limited research on task-related brain activity in older adults. The current study administered a naturalistic version of the TMT to a healthy older-aged population in an MRI environment using a novel, MRI-compatible tablet. Functional MRI was conducted during task completion, allowing characterization of the brain activity associated with the TMT. Performance on the TMT was evaluated using number of errors and seconds per completion of each link. Results are reported for 36 cognitively healthy older adults between the ages of 52 and 85. Task-related activation was observed in extensive regions of the bilateral frontal, parietal, temporal and occipital lobes as well as key motor areas. Increased age was associated with reduced brain activity and worse task performance. Specifically, older age was correlated with decreased task-related activity in the bilateral occipital, temporal and parietal lobes. These results suggest that healthy older aging significantly affects brain function during the TMT, which consequently may result in performance decrements. The current study reveals the brain activation patterns underlying TMT performance in a healthy older aging population, which functions as an important, clinically-relevant control to compare to pathological aging in future investigations.
Citation: Talwar N, Churchill NW, Hird MA, Tam F, Graham SJ, Schweizer TA (2020) Functional magnetic resonance imaging of the trail-making test in older adults. PLoS ONE 15(5): e0232469. https://doi.org/10.1371/journal.pone.0232469
Editor: Thalia Fernandez, Universidad Nacional Autonoma de Mexico, MEXICO
Received: August 16, 2019; Accepted: April 15, 2020; Published: May 12, 2020
Copyright: © 2020 Talwar et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: We have also made the dataset available on FigShare (DOI: 10.6084/m9.figshare.12156645, link: https://figshare.com/articles/TMT_HealthyAging_fMRIData/12156645).
Funding: The work completed in the current study was supported by funding provided to T.A.S. from the Ministry of Transportation Ontario through the Road Safety Community Partnership Program (https://www.ontario.ca/page/ministry-transportation) and an Alzheimer's Association Research Grant (AARG-16-442685) (https://www.alz.org/). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
The trail-making test (TMT) is a pen-and-paper neuropsychological test that is used extensively in research and clinical settings. The TMT consists of two different conditions: TMT-A and TMT-B, each to be performed as quickly as possible while maintaining accuracy, and without removing the tip of the pen from the page[1]. For TMT-A, individuals must connect 25 randomly distributed encircled numbers from 1 to 25 in ascending order (i.e. 1-2-3-…). For TMT-B, individuals must connect numbers and letters in alternating, ascending order (i.e. 1-A-2-B-3-C…). The task conditions are scored using metrics that relate to completion times and the number of errors committed. Successful performance of the TMT requires the engagement of numerous cognitive elements including visual search, attention, processing speed, task switching, cognitive flexibility and executive function [1–3].
The sensitivity of the TMT to dysfunction in many cognitive domains has led to its widespread use as a clinical tool to measure cognitive impairment associated with various neurological and psychiatric conditions [4–12]. For example, the TMT is used in clinical settings to screen for neurodegenerative disease in older adults, such as Alzheimer’s Disease (AD) [4,5,7,9–11]. Much of the research investigating the TMT has focused on comparing different performance metrics and measuring changes in TMT performance for elderly and cognitively impaired populations [7,9–11,13,14]. Prior research has established that TMT behavioural performance is significantly impacted by healthy aging, with older adults exhibiting longer task completion times [13–16], likely arising from age-related decline in cognitive domains such as attention, processing speed and mental flexibility [15,16].
At present, however, there are insufficient functional neuroimaging data available for the TMT to increase our scientific understanding beyond this rather general statement. One promising tool for evaluating the neural mechanisms underpinning TMT performance is functional magnetic resonance imaging (fMRI), which measures task-related brain activity based on the blood-oxygenation-level-dependent (BOLD) effect of neurovascular coupling [17]. The fMRI method enables non-invasive measurement of activity throughout the brain volume at spatial and temporal resolutions of millimetres and seconds, respectively. Early fMRI studies of TMT performance involved highly modified versions of the task, with the drawing component replaced by a verbal response or a button-press response [18–20]. Although these studies provided insight into some of the cognitive elements that are required for successful TMT performance, use of a simplified response method meant that it was impossible to characterize the specific brain regions involved in sensation, cognition and action accurately and in their entirety[18–20].
To overcome this limitation and study the TMT in more detail, specialized MRI-compatible technology was developed that facilitated drawing responses (and writing responses) during fMRI. This technology started with a fiber-optic MRI compatible writing device [21] and progressed to a computerized MRI-compatible tablet system with a touch-sensitive screen [22]. Early prototypes of the tablet system were used to deliver simplified versions of the TMT during fMRI [21–23]. However, participants were unable to see their hand, the tablet or the stylus while they were lying supine in the magnet bore engaged in task performance, resulting in their TMT responses relying heavily on proprioception. In the clinical setting, the patient or research participant sits upright with full visualization of the page and their writing movements. The reliance on proprioceptive input sparked further development of the tablet technology to include a video camera and augmented reality display to provide real-time visual feedback of hand position during task completion [24]. Recently, this new tablet prototype was validated as an effective, naturalistic method for administering the TMT and brain activity was investigated during TMT performance in a cohort of healthy young adults [25].
Although functional neuroimaging of TMT performance in healthy young adults is an important first step [18–21,25,26], to our knowledge, fMRI studies of the task have yet to be performed in older individuals–the population where the TMT is predominantly administered in the clinic. The cognitive changes that occur during healthy aging are well known to be accompanied by changes in brain activity, which can present on average as decreased activity associated with functional decline as well as increased activity associated with neural compensation mechanisms [27,28]. Consequently, it is necessary to study the neural correlates of TMT performance in healthy older adults from a basic science perspective, to better understand the effects of cognitive aging, as well as to provide initial normative data prior to conducting clinically-focused fMRI studies of the TMT in elderly patient populations. The present fMRI study addresses these goals by studying the brain and behavioural relationships generated by administering TMT-A and TMT-B to healthy adults over the age of 50, using the enhanced tablet system. It is hypothesized that both task conditions (TMT-A and TMT-B) will recruit extensive brain areas in the bilateral frontal and parietal lobes across the older aging group, consistent with results found in previous fMRI studies with younger populations. Based on prior literature examining TMT performance in healthy older adults [13–16], it is hypothesized that there will be subtle behavioural performance declines in in-scanner TMT performance as represented by slower task completion times, with a corresponding reduction in amplitude and spatial extent of activation within TMT-related brain areas. Age-related declines in activation are further hypothesized to be predominantly in frontal areas involved in TMT performance. Frontal regions, which are essential for controlling cognitive processes involved in the TMT, are susceptible to the effects of age and, therefore, are a likely neural correlate of worsened task performance observed in older adults.
Materials and methods
Participants
Thirty-seven (n = 37) healthy participants between the ages of 52 and 85 were recruited from the local community into this study. The participants were all screened for history of any neurological or psychiatric condition, neurological incident or substance abuse. The participants were all right handed as determined by the Edinburgh Handedness Inventory [29] and were fluent in English. The study was approved by the Research Ethics Board at St. Michael’s Hospital according to the principles of the Declaration of Helsinki. Participants provided their free and informed consent prior to entering the study.
Neurocognitive assessment
The Montreal Cognitive Assessment (MoCA) [30] was used to evaluate the cognitive abilities of the participants prior to fMRI. The MoCA score was used to determine whether there were any subjects with significant cognitive impairments. The inclusion criterion was that all participants were to have a MoCA score within 2.5 standard deviations of the group mean. One subject was identified as an outlier on this basis (MoCA score = 15, p ~ 0.01, Normal distribution, 2-tailed) and was not included in any analyses. Participants also completed one sample trial and one full trial of both TMT-A and TMT-B on paper, as the test is normally administered. The paper version of the TMT was administered to investigate convergent validity with the tablet-based TMT. Participants were instructed to work as quickly as they could without making mistakes and not to lift the pen from the paper. The number of errors, number of pen lifts and total time to complete the task was recorded for both the sample and full trials. No time limitation was imposed on performing TMT-A or TMT-B.
Magnetic resonance imaging
Participants were imaged using the St. Michael’s Hospital Research 3.0 Tesla MRI system (Magnetom Skyra, Siemens Healthineers, Erlangen, Germany), with the standard 20-channel head coil. The structural imaging was performed using the T1-weighted Magnetization Prepared Rapid Acquisition Gradient Echo method (MPRAGE: inversion time (T1)/echo time (TE)/repetition time (TR) = 1090/3.55/2300 ms, flip angle (FA) = 80, bandwidth (BW) = 200 Hz/pixel, sagittal orientation with field of view (FOV) = 240x240 mm, 256x256 matrix, 192 slices, voxels = 0.9 mm x 0.9 mm x 0.9 mm). The functional MRI data were acquired during completion of the TMT using multi-slice T2*-weighted echo planar imaging (EPI: TE/TR = 30/2000 ms, FA = 700, BW = 2298 Hz/px, oblique-axial slices, slices interleaved ascending, with FOV = 200x200 mm, 64x64 matrix, 32 slices with 4.0 mm thickness and 0.5 mm gap, voxels = 3.125 mm x 3.125 mm x 4.5 mm).
MRI-compatible tablet system
The MRI-compatible tablet system parallels real-world pen and paper settings by using a touch screen and stylus in conjunction with an augmented reality system to provide real-time visual feedback of hand and stylus position [24]. The tablet and camera were mounted on an adjustable stand, which was placed over the waist of the participant during the imaging session. A mirror was placed on the top of the head coil and angled towards a screen, on which the task and augmented reality environment were displayed using an MRI-compatible projector (Avotec, Stuart, FI). Pads were placed underneath the elbows of the participant to provide comfort and limit movement during task performance. MRI-compatible prescription glasses (MediGlasses for fMRI, Cambridge Research Systems, Kent, UK) were used for participants with corrected vision. The experimental set-up is shown in Fig 1.
(A) The participant lies supine on the MRI table before entering the magnet, with the tablet mounted at the waist using an adjustable stand. The head coil for receiving fMRI signals includes a mirror to view the task in an augmented reality environment on a display screen using an fMRI-compatible projector (not shown). (B) The tablet mount includes an fMRI-compatible video camera for a top-down view of tablet interactions. (C) Image results of the participant completing the TMT-B task on the tablet. TMT = Trail Making Test. fMRI = functional magnetic resonance imaging.
fMRI of the TMT
The tablet-based TMT was developed and presented using E-Prime Version 2.0 (Psychology Software Tools, Sharpsburg, PA). Outside of the MRI system, participants were familiarized with the equipment by completing a training session with the tablet and by viewing the set-up in the magnet room. Once in the magnet bore, participants completed additional tablet training to become comfortable with the set-up by performing simple tasks such as writing their name or drawing a house. Participants subsequently completed the tablet-based TMT according to the timing diagram shown in Fig 2, after structural MRI was performed. Two “runs” of fMRI time series data were collected, with each run consisting of TMT-A (40 s) and TMT-B (60 s) interleaved with a visual fixation condition (10 s). Two “trials” of each condition were performed within each run. At the beginning of each run, the participants received on-screen instructions on how to perform each task. Four different iterations with pseudorandom patterns of numbers and letters for both the TMT-A and TMT-B were used overall, two in each run. Both versions of the TMT had 25 targets to link in each iteration. Specifically all trials of TMT-A had 25 numbers (1–25) and all trials of TMT-B had 13 numbers (1–13) and 12 letters (A-L). For the fixation task, participants were asked to fixate their eyes on the center of a black cross on a white background. This version of the tablet-based TMT has been used previously in research and has been validated as a realistic replication of the paper-based TMT [22,25].
TMT = Trail Making Test.
Data analysis
Analysis of behavioural measures.
The traditional paper-based TMT delivery adopts the completion time as a key measure of performance. However, as is typical of a “block-design” fMRI experiment, brain activity associated with tablet-based TMT delivery was assessed for a fixed time duration. Based on previous tablet-based fMRI of the TMT in young adults [25], it was predicted that some participants would not complete TMT-A and/or TMT-B in the allotted time duration. Given that this precluded measurement of completion time across all participants, a different normalized metric was adopted: the seconds per link (SPL). The SPL value was calculated for both paper-based and tablet-based delivery by dividing the total time spent performing one part of the TMT (i.e. TMT-A or TMT-B) by the number of correct links completed during that time. Paper-based and tablet-based TMT performance was also compared using a secondary measure, the number of errors (NE). The NE value included both self-corrected and uncorrected errors that occurred during the allotted time duration. Given that errors were rare (see Results), no normalization was incorporated into the comparison.
In preliminary behavioural analysis, the effects of trial on SPL were evaluated using a 1-way repeated measures ANOVA, where a significant effect was found (F = 3.94, p = 0.009). Post-hoc statistical testing using Tukey’s test determined that the first tablet-based trial of both TMT-A and TMT-B for run 1 was significantly different from the rest of the tablet-based trials (Trial 2–1, p = 0.03, Trial 3–1, p = 0.02, Trial 4–1, p = 0.78), whereas there was no significant difference between the remaining trials (Trial 3–2, p = 1.00, Trial 4–2, p = 0.26, Trial 4–3, p = 0.23). During the first trial, participants may have been acclimatizing to the fMRI environment and performance of the tablet-based TMT, leading to slower initial performance. The behavioural measures were averaged for each participant across the other three trials of TMT-A and three trials of TMT-B, to improve the statistical power to detect significant behaviour and brain activation effects. To measure the effect of both versions of the task (TMT(A+B)), the behavioural measures were averaged across all six trials of both TMT types. To isolate the effect of TMT-B, the behavioural results for TMT-A were subtracted from the results of TMT-B for each performance measure (TMT-(B-A)) in analogous fashion. Average task performance was correlated with age using a Spearman’s rank correlation. To provide convergent validity of the tablet-based TMT, the SPL for the tablet-based TMT-A, TMT-B and TMT-(B-A) were correlated with the results from the paper version using a Spearman’s rank correlation. Additionally, the SPL for TMT-A, TMT-B and TMT-(B-A) were compared between the two versions of the task (i.e. paper and tablet) using a paired Mann-Whitney U test.
fMRI preprocessing and analysis.
The fMRI data and structural scans were visually inspected for any gross abnormalities. A hybrid pipeline, which includes tools from the Analysis of Functional Neuroimages (AFNI; https://afni.nimh.nih.gov) package [31], the FMRIB Software Library (FSL; https://www.fmrib.ox.ac.uk/fsl) package [32] and algorithms established in the laboratory, was used for preprocessing and analysis of the imaging data. The preprocessing pipeline included the following sequence of steps: rigid-body motion correction (AFNI 3dvolreg), slice-timing correction (AFNI 3dTshift), removal of outlier scan volumes via SPIKECOR [33], spatial smoothing with an isotropic Gaussian kernel of 6 mm full width at half maximum (AFNI 3dmerge), followed by regression of motion parameters as well as linear and quadratic trends. PHYCAA+ [34] was used to perform data-driven down-weighting of signal from vascular and ventricular regions, followed by further noise reduction using white matter signal regression from seed regions of interest placed in the left and right corona radiata (average cluster size = 1052 mm3, Montreal Neurological Institute (MNI) coordinates, left side = -30, -8 24, right side = 26, -8, 24) and the left and right lateral ventricles (average cluster size = 440 mm3, MNI coordinates, left side = -10, -14, 22, right side = 8, -14, 22).
Following preprocessing, brain images were transformed into a common anatomical template space as follows: the FSL flirt algorithm was used to calculate the rigid-body transform of the mean fMRI brain volume to the T1-weighted anatomical brain volume, and the affine warp of the T1-weighted anatomical brain volume for each participant to the MNI152 (Montreal Neurological Institute) template [35] for each participant. The net transformation matrices were applied to the fMRI imaging data, which were resampled at a resolution of 3 mm x 3 mm x 3 mm. Due to the variability in brain size and morphology amongst the older adults (80+ years-old), the anatomical images were all visually inspected after both segmentation and transformation steps, and manually adjusted in cases where errors were identified.
The first level of analysis on the preprocessed imaging data was completed at the individual participant level using the NPAIRS analysis framework [36]. The BOLD fMRI signal response was estimated using a general linear model (GLM) to fit each task contrast of interest after convolving the design matrix with a canonical double-gamma hemodynamic response function [37], thereby obtaining a map of voxel-wise beta coefficient estimates. This was done separately for each task (TMT-A and TMT-B) run on the three trials of interest (i.e. ignoring trial one), then the corresponding pair of beta coefficient maps were used to obtain reproducible, z-scored maps of reliable task-related activation for each participant using the split-half statistical procedure of Strother et al. [36]. The analyses were done separately for two task contrasts, one which isolated activation associated with performing both parts of the TMT (i.e. TMT-(A+B) minus fixation) and another which isolated the different neural demands of TMT-B in relation to TMT-A (TMT-(B-A)).
The initial group analysis identified mean group-level activation maps based on the z-scored individual participant maps, using a one-sample t-test to identify significant areas of group-related activation. Significance was reported at a nominal voxel-level threshold of p<0.005, with cluster-size thresholding to control for false positives (cluster size = 253 for TMT-(A+B)-fixation; cluster size = 227 for TMT-(B-A)). Cluster-size thresholding was performed using a non-parametric permutation approach. For each analysis, activation maps were thresholded at a nominal p = 0.005 level. To obtain a corresponding null distribution, labels were randomly permuted, the data was re-analyzed and the same voxel-level threshold (2000 iterations) was applied. From the resulting data, the minimum cluster size (cmin) was obtained such that the cumulative probability p(c>cmin) = 0.05, and subsequently retained all cluster of size c>cmin. Additional regression analyses tested for an association between individual participant activation z-scores and covariates of interest, including age and SPL. This was done by calculating a t-statistic on the regression coefficient to identify brain regions with significant loadings, with thresholding and effect size reporting done as described above. The group-level analyses were performed separately for the TMT-(A+B)-fixation and TMT-(B-A) contrasts. The SPL used in the covariate analysis was an average of both the TMT-A SPL and TMT-B SPL for the TMT-(A+B)-fixation contrast, and was TMT-A SPL subtracted from TMT-B SPL for the TMT-(B-A) contrast. Given the novelty of the data and potential importance in guiding future age-related research, for non-significant findings an exploratory analysis was conducted at the previously established voxel-level threshold of p = 0.005, but a relaxed cluster-size threshold of c = 20.
Motion parameters were reported for all participants as the root-mean-square (RMS) of displacement in 3D space. This metric was correlated with age and performance on the TMT using Spearman’s correlation.
Results
Participants
Thirty-six participants were included in the final analysis. One participant was excluded from analysis due to an outlying MoCA score. Demographic and neuropsychological data for the participants are summarized in Table 1. The cohort had a median age of 78 years (inter-quartile range: 14.5 years); had similar proportions of females and males; and was well-educated. Participants performed similarly on neuropsychological testing with small interquartile ranges for the MoCA and TMT-A, but with more variability in TMT-B and TMT-(B-A) performance.
fMRI behavioural results
The results of the tablet-based TMT showed a statistically significant correlation between age and various behavioural metrics for TMT-A, TMT-B, the average of both (denoted TMT-(A+B)) and the difference between TMT-B and TMT-A (denoted TMT-(B-A); see Table 2). For TMT-A, there was a significant positive correlation between age and SPL (rho = 0.594, p < 0.001). For TMT-B, there was a significant positive correlation between age and SPL (rho = 0.612, p < 0.001) and NE (rho = 0.404, p = 0.014). For TMT-(A+B), there was a significant positive correlation between age and SPL (rho = 0.647, p < 0.001) and NE (rho = 0.360, p = 0.031). Similarly, for TMT-(B-A) there was a significant positive correlation between age and SPL (rho = 0.439, p = 0.007) and NE (rho = 0.406, p = 0.014).
In addition, the performance on the tablet-based TMT tasks was significantly correlated with the results observed on the paper versions. Both TMT-A and TMT-B had a significant positive correlation between SPL on the two versions of the task with moderate strength (TMT-A, rho = 0.486, p < 0.005; TMT-B, rho = 0.405, p = 0.015). However, there was a weak, non-significant correlation between the two task versions for SPL of TMT-(B-A) (rho = 0.142, p = 0.409).
There was a significant difference in task performance between the tablet and paper versions of the TMT. For both TMT-A and TMT-B, there was a significant difference in SPL (TMT-A, p < 0.001; TMT-B, p = 0.046). Meanwhile for TMT-(B-A), the difference in SPL was not significant (p = 0.143).
fMRI results
Fig 3A displays brain regions with significant group-level activation for the contrast of TMT-(A+B) with fixation, with significant clusters reported in Table 3. A spatially extensive pattern of positive activity was observed in the brain (i.e., increased activity during TMT-(A+B) compared to fixation), with positive activations seen in the visual cortex that extend into both the parietal lobes and upper cerebellum. Activity was also observed in prefrontal regions, centered mainly in regions of the inferior and middle frontal gyri, sensorimotor areas and the supplementary motor area, also encompassing subcortical areas including the insula, putamen and pallidum. Negative group-level activations (i.e. decreased activity during TMT-(A+B) compared to fixation) were also spatially extensive, concentrated predominately in areas of the posterior cingulate and angular gyri as well as the bilateral temporal poles, hippocampal formation and superior frontal lobe.
Group-level activation maps for tablet-based performance of (A) the TMT-(A+B) contrasted with fixation, and (B) TMT-B contrasted with TMT-A (i.e. TMT-(B-A)).
Fig 3B displays significant group-level brain activity associated with the TMT-(B-A) contrast, with significant clusters reported in Table 3. Positive activity (i.e., increased activity in TMT-B compared to TMT-A) was observed in various cortical areas, predominantly lateralized to the left hemisphere. Positive activation was observed in the left inferior parietal lobe, including the angular and supramarginal gyri, as well as the inferior, middle and superior gyri of the left frontal lobes. Additionally, the left insular cortex and middle cingulate gyrus exhibited significant clusters of positive activity. Negative activations (i.e., decreased activity in TMT-B compared to TMT-A) were focused in posterior brain regions, namely the bilateral cerebellum and occipital lobes, extending into areas of the middle temporal lobe and the precuneus of the parietal lobe.
Fig 4 identifies brain areas that show significant effects of age on brain activity for the contrast of TMT-(A+B) with fixation, with significant clusters reported in Table 4. There was no significant positive association between brain activity and age. A distributed pattern of brain regions showed negative associations of brain activity with age, with large clusters in the bilateral occipital lobes, middle temporal lobes and cerebellum as well as small clusters in regions of the right parietal lobes, including the postcentral gyri and the inferior parietal lobe. There were no significant associations between age and brain activity in the TMT-(B-A) contrast.
Regions that show significant associations between age and brain activity for tablet-base performance of (A) TMT-(A+B) contrasted with fixation.
No brain regions had a statistically significant associations with our measure of task performance (SPL), for both contrasts (TMT-(A+B) vs fixation and TMT-(B-A)). A follow-up exploratory analysis was therefore completed at a relaxed cluster-size threshold (see Table 5 for cluster report). The results identified additional brain regions that had negative associations of brain activity with age in the TMT-(A+B) vs. fixation contrast, including bilateral regions of the thalamus and precentral gyrus, left lateralized regions of the parietal lobe (postcentral gyrus, superior parietal lobe), lingual gyrus and supplementary motor area and right lateralized areas of the frontal lobe (inferior and middle frontal gyrus) and supramarginal gyrus. Two clusters also showed positive associations of brain activity with age, centered in the right rectus and the right medial frontal gyrus (pars orbitalis). Several brain areas exhibited positive associations with age in the TMT-(B-A) contrast, including larger clusters located in the left cerebellum and occipital lobe and small clusters in the left temporal lobe, right insula and bilateral frontal lobes (left superior frontal gyrus, right middle frontal gyrus).
In the exploratory analysis, performance (SPL) had both negative and positive associations with brain activity in the TMT-(A+B) vs. fixation contrast. Clusters of negative effects were centered on the left hemisphere in regions of the cerebellum and the superior frontal gyrus. Meanwhile, clusters of positive effects were located bilaterally in the right inferior temporal gyrus as well as the left caudate nucleus and left superior occipital gyrus.
Similarly, brain activity in the TMT-(B-A) contrast showed both negative and positive associations with performance (SPL). Negative effects were observed bilaterally in the cerebellum, parahippocampal gyri and inferior frontal gyri and left lateralized in the occipital and temporal lobe. Positive effects were centered in the bilateral caudate nuclei, the right frontal lobe and motor areas, and the left cerebellum and occipital lobe.
The average motion parameters across both fMRI scans (Run 1 and Run 2) are listed for each participant in S1 Table. Head motion showed a moderate, significant correlation with age (rho = 0.439, p = 0.007) and a weak, significant correlation with the performance metrics for TMT-B (SPL, rho = 0.334, p = 0.046; NE, rho = 0.331, p = 0.049).
Discussion
This study was the first to investigate brain activity during TMT performance in a cohort of older adults, using novel fMRI-compatible tablet technology to administer a realistic replication of the TMT in an MRI environment. All participants went through extensive screening to ensure they were cognitively intact and the group median score on the MoCA was 27, which is classified as a normal, healthy score [30]. For the group, contrasting both conditions of the TMT (TMT-A and TMT-B) with fixation revealed extensive bilateral recruitment of brain areas in the frontal, parietal, temporal and occipital lobes and key motor areas common to both tasks. In comparison, contrasting TMT-B with TMT-A identified task-related activity in the frontal and parietal lobes that was largely lateralized to the left hemisphere. Within this group, aging was significantly associated with brain activity for the contrast of TMT-B and TMT-A with fixation. For this contrast, older age was significantly associated with an extensive decrease in brain activity, particularly in regions of the bilateral occipital, temporal and parietal lobes, as well as worse performance on the task.
The activation patterns found in this study are generally consistent with those found in previous imaging studies of the TMT for younger populations (~20–35 years old) [18,19,21,22,25,38,39]. There were some slight differences with other studies, which are elaborated on later in the discussion. In contrast to fixation, both TMT conditions were associated with activation of the bilateral frontal, parietal, temporal and occipital lobes. The frontal lobe and more specifically the prefrontal cortex, has been historically associated with the TMT due to its role in executive functions, such as attention, processing speed and planning, which are crucial for task coordination and completion [18,19,21,22,25,26,38–40]. The medial temporal lobe has also been linked to executive functions involved in the TMT [41–43]. In addition, the temporal lobe may be recruited due to working memory demands of recalling the numbers and letters during the TMT [21,22]. Both the occipital and parietal lobes have been implicated in TMT performance due to their involvement in visual search abilities [25,44,45]. Specifically, the parietal lobe aids in task completion through its role in visuospatial ability, which helps integrate visual information to guide movements in space [46,47].
Contrasting TMT-B to TMT-A revealed positive activation that was largely lateralized to the left hemisphere for the frontal and parietal lobes. Left lateralization of brain activity has been previously observed in neuroimaging studies of TMT-B compared to TMT-A [18,21,22,25]. However, in a study by Mueller et al., it was identified that TMT-B was found to evoke bilateral activity in prefrontal and premotor areas, with more activity observed in the right hemisphere [39]. Similarly, Shibuya-Tayoshi et al. found that the majority of their participants displayed bilateral patterns of activation in the prefrontal cortex during completion of TMT-B [26]. Both aforementioned studies used functional near-infrared spectroscopy (fNIRS) to measure brain activity, meanwhile studies employing fMRI, this study included, have observed a strong left hemisphere lateralization of brain activity associated with TMT-B. The different neuroimaging modalities and resulting methods of TMT delivery may contribute to the discrepancy in results between these studies.
Investigations of the TMT in younger populations have reported increased activity in the frontal, parietal and temporal lobes during TMT-B compared to TMT-A, consistent with what was found in this investigation [22,25]. As previously noted, these cortical areas have important cognitive functions for TMT completion and may be more active during TMT-B due to the increased complexity of this condition. Negative task-related activity in both contrasts was observed in areas involved in the default mode network (DMN), including the posterior cingulate cortex, precuneus, angular gyri, hippocampus, inferior parietal lobe and temporal lobe [48–50]. Previous research supported DMN activation during periods of wakeful rest [51] and DMN suppression during cognitive stimulation, including performance of complex tasks (e.g. TMT).
Two previous fNIRS studies also investigated brain activity associated with the TMT in an older population (50+ years old) and compared the results between young and old age groups [38,39]. Both observed more activated channels in the older group when contrasting TMT-B with TMT-A [38,39], similar to the results found in this study. However, as an imaging method, fNIRS has restricted spatial resolution and can only measure activity 1.5 cm deep into the cortex of the brain, which can limit precision when identifying activated brain areas [52]. Functional MRI is advantageous in comparison because activity is measured with millimetre spatial resolution throughout the entire brain volume, as appropriate for studying the TMT. For example, a recent study by Churchill et al. (2017) used an earlier version of the MRI-compatible tablet technology [22] to deliver a modified version of the TMT to young (21–33 years old) and older (61–82 years old) cohorts [53]. The older cohort showed less brain activity in posterior brain areas, supporting the results of the current investigation.
One limitation of these previous fNIRS and fMRI studies was that the experiments were designed to compare the average brain activity between two different age cohorts. Such an approach may not account for age-related variation in brain activity within each cohort, especially if the age-range within a given cohort is substantial; this has been previously observed for TMT-related brain networks, e.g., in (Churchill et al. (2017)). This effect was observed in the present study, which showed that there were performance and brain activity differences between participants at the younger end of the cohort (e.g. 60 years old) and those at the older end (e.g. 75 years and older). Thus, examining age-related variability in brain function as a continuous covariate may provide additional useful information.
The current study revealed extensive areas across the bilateral occipital, middle temporal and parietal lobes with reduced activity in older aged participants during both TMT conditions contrasted to fixation. Increased age was associated with longer time to complete links and more errors during the task, which is consistent with current literature on the age-related changes in TMT performance [13–16]. Previously it was suggested that cognitive deficits in processing speed, attention and mental flexibility may be affecting performance in healthy, elderly populations [15,16]. The current findings support existing literature, which identifies age-related cognitive slowing and decline in many domains including memory, spatial ability, attention, processing speed, reasoning and global cognition, irrespective of health and education levels [54–56]. This is in accordance with previous studies, which have revealed both performance decrements and reduced levels of brain activity to be associated with increased age in a cognitively healthy older adult population [57,58]. Importantly, the results of this study clarify the neural basis underlying specific age-related changes in task performance, since the identified brain areas are implicated in cognitive functions vital for the TMT. The occipital, temporal and parietal lobes have well-characterized age-related changes in structure and function, which affect cognitive abilities of the elderly population in many domains including attention, visual processing, executive function and working memory [59–67]. The medial temporal lobe in particular is of substantial interest aging studies, due to its role as a clinical biomarker for forms of dementia [41,68–73].
Contrasting TMT-B to TMT-A revealed non-significant effect when adjusted for multiple comparisons. However, a more relaxed threshold identified greatest effects on brain activity in areas of the left cerebellum and occipital lobe associated with increased age. Due to the increased difficulty of TMT-B, as demonstrated by the higher SPL values, greater cognitive resources are required to complete the task and this effect may be exacerbated in older adults who need to exert more cognitive effort for task completion. Older participants had higher activity in the occipital lobe, which has identified visual processing and visual search functions, and would be necessary for localizing numbers and letters throughout task completion. The cerebellum has been identified as important for executive functions, particularly involving multitasking, working memory and divided attention [74–76], which would be increasingly important for TMT-B completion as it requires switching between numbers and letters. Behaviourally, older age was significantly associated with increased errors and slower completion (SPL during TMT-B). This suggests that despite increased brain activity during TMT-B, older participants tend to perform worse on the task. The age-related effect on brain activity was more extensive for the TMT-(A+B) vs fixation contrast. This may be because the TMT-(B-A) contrast has more focal patterns of activation, therefore there are fewer brain areas where an aging affect can be observed.
Sparse patterns of activation were correlated with task performance as measured by SPL of TMT-(B-A). When contrasting TMT-B with TMT-A, decreased activity in the left inferior occipital lobes and bilateral cerebellum was associated with worse performance (i.e. increased SPL). As previously discussed, the occipital lobes and cerebellum have significant cognitive functions that are involved in task completion. Research studies have established that increased neural efficiency is associated with better cognitive performance [77–79], suggesting that skilled participants have more efficient brain functioning leading to decreased brain activity and recruitment of less brain regions.
The results of this study reveal the neural networks underlying TMT performance in a healthy older adult population–a clinically relevant control population due to the widespread use of the TMT as a screening tool for dementia. With the present results serving as the foundation, future fMRI research should investigate brain activity during TMT performance in patients with mild cognitive impairment (MCI) as well as those with early probable AD. Such research could provide valuable information on TMT-related brain activity in the full spectrum of cognitive impairment, and on how the associated activation patterns relate to the observed deficits in task performance. In addition, such research will help to inform clinicians on the neuropathology detected by the TMT and may potentially reveal new combined behavioural and neuroimaging biomarkers that characterize impaired TMT performance.
The current study uses novel technology to replicate a cognitive test environment using the best methods to date. However, it is not an exact replication with a few limitations to be noted, including the supine position of participants, the limited field of view and restricted task block lengths. Participants were allotted 60 seconds to complete TMT-B to align with a block-design fMRI experiment, which focuses on networks that are most consistently engaged across participants and over time. This design may limit the ability to observe neurocognitive processes, which emerge later in the TMT task specifically among individuals who complete the task slower. More detailed future analyses should be conducted to identify transient components of the task with a different task block design. Additionally, the tablet may have added difficulty to completion of the TMT as shown by the significant difference in SPL between tablet and paper versions of the TMT. Although there was a significant correlation in SPL between the two versions for TMT-A and TMT-B, there was no significant correlation for TMT-(B-A). This is a potential limitation of the tablet version of the TMT and should be addressed in future studies to ensure it is more consistent with the traditional paper version.
As a final note, it is important to recognize that the physiological effects of aging on the BOLD fMRI signal are a potential confound to interpreting the study findings [80]. Future studies should incorporate additional physiological imaging measures, such as BOLD fMRI of cerebrovascular reactivity during a breath-hold task, to control for neurovascular coupling confounds [81] and confirm results found in this study. Notwithstanding this recommendation, the present results are in good agreement with prior research investigating the brain activation patterns underlying the TMT, suggesting that the age-related physiological confounds have not significantly impacted the results of this study.
Conclusions
Using novel tablet technology and fMRI, this was the first study to investigate neural areas recruited during a realistic version of the TMT in an older adult population (50+ years-old). Across the older aging group, both conditions of the TMT recruited activity from extensive brain regions across the bilateral frontal, parietal, temporal and occipital lobes. Older age had an impact on brain activity and function during both TMT conditions, such that older age was associated with reduced brain activity and worsened task performance. In particular, there was decreased task-related activity in the medial temporal lobe, which is a common clinical biomarker for dementia. This finding is important as it provides additional support for the use of the TMT as a clinical screening tool for dementia. Future studies are necessary to determine the changes in brain activity patterns in cohorts with diagnosed cognitive impairment along the spectrum from mild cases (MCI) to more severe cases (AD). This will allow us to determine the difference between task-related neural networks employed in a healthy aging and a cognitively impaired population, providing insight into the neuropathology underlying impaired TMT performance and the efficacy of the TMT as a clinical screening tool.
Supporting information
S1 Table. The average motion parameters for each participant.
Motion parameters for each participant averaged across run 1 and 2.
https://doi.org/10.1371/journal.pone.0232469.s001
(DOCX)
Acknowledgments
The authors thank Tahira Tasneem, Breanna Jessop and Iryna Pshonyak for their help running fMRI scans as well as the MRI Research Imaging Department at St. Michael’s Hospital.
References
- 1.
Lezak MD, Howieson DB, Loring DW, Hannay HJ, Fischer JS. Neuropsychological Assessment. 4th ed. New York, NY, US: Oxford University Press; 2004.
- 2. Salthouse TA. What cognitive abilities are involved in trail-making performance? Intelligence [Internet]. NIH Public Access; 2011 Jul [cited 2018 Jul 19];39(4):222–32. Available from: http://www.ncbi.nlm.nih.gov/pubmed/21789028 pmid:21789028
- 3. Kortte KB, Horner MD, Windham WK. The trail making test, Part B: Cognitive flexibility or ability to maintain set? Appl Neuropsychol [Internet]. 2002 Jun [cited 2018 Nov 15];9(2):106–9. Available from: http://www.ncbi.nlm.nih.gov/pubmed/12214820 pmid:12214820
- 4. Auriacombe S, Orgogozo J-M, Dartigues J-F, Fabrigoule C, Amieva H, Lafont S, et al. Analysis of Error Types in the Trail Making Test Evidences an Inhibitory Deficit in Dementia of the Alzheimer Type. J Clin Exp Neuropsychol (Neuropsychology, Dev Cogn Sect A) [Internet]. Taylor & Francis Group; 1998 Apr 9 [cited 2018 Nov 15];20(2):280–5. Available from: http://www.tandfonline.com/doi/abs/10.1076/jcen.20.2.280.1161
- 5. Morris JC, Heyman A, Mohs RC, Hughes JP, van Belle G, Fillenbaum G, et al. The Consortium to Establish a Registry for Alzheimer’s Disease (CERAD). Part I. Clinical and neuropsychological assessment of Alzheimer’s disease. Neurology [Internet]. 1989 Sep [cited 2018 Nov 15];39(9):1159–65. Available from: http://www.ncbi.nlm.nih.gov/pubmed/2771064 pmid:2771064
- 6.
Greenlief C, Margolis RB, Erker GJ. Application of the Trail Making Test in Differentiating Neuropsychological Impairment of Elderly Persons. Percept Mot Skills [Internet]. SAGE PublicationsSage CA: Los Angeles, CA; 1985 Dec 31 [cited 2018 Nov 15];61(3_suppl):1283–9. Available from: http://journals.sagepub.com/doi/10.2466/pms.1985.61.3f.1283
- 7. Ashendorf L, Jefferson AL, O’connor MK, Chaisson C, Green RC, Stern RA. Trail Making Test errors in normal aging, mild cognitive impairment, and dementia. Arch Clin Neuropsychol [Internet]. 2008 [cited 2018 Sep 10];23(2):129–37. Available from: https://academic.oup.com/acn/article-abstract/23/2/129/2934 pmid:18178372
- 8. Beckham JC, Crawford AL, Feldman ME. Trail making test performance in Vietnam combat veterans with and without posttraumatic stress disorder. J Trauma Stress [Internet]. Kluwer Academic Publishers-Plenum Publishers; 1998 Oct [cited 2018 Nov 15];11(4):811–9. Available from: http://doi.wiley.com/10.1023/A%3A1024409903617 pmid:9870231
- 9. O’Rourke JJF, Beglinger LJ, Smith MM, Mills J, Moser DJ, Rowe KC, et al. The Trail Making Test in prodromal Huntington disease: contributions of disease progression to test performance. J Clin Exp Neuropsychol [Internet]. NIH Public Access; 2011 Jun [cited 2018 Nov 15];33(5):567–79. Available from: http://www.ncbi.nlm.nih.gov/pubmed/21302170 pmid:21302170
- 10. Weintraub D, Moberg PJ, Culbertson WC, Duda JE, Katz IR, Stern MB. Dimensions of Executive Function in Parkinson’s Disease. Dement Geriatr Cogn Disord [Internet]. 2005 [cited 2018 Nov 15];20(2–3):140–4. Available from: http://www.ncbi.nlm.nih.gov/pubmed/16020942 pmid:16020942
- 11. Bossers WJR, van der Woude LH V, Boersma F, Scherder EJA, van Heuvelen MJG. Recommended measures for the assessment of cognitive and physical performance in older patients with dementia: a systematic review. Dement Geriatr Cogn Dis Extra [Internet]. Karger Publishers; 2012 Jan [cited 2018 Nov 15];2(1):589–609. Available from: http://www.ncbi.nlm.nih.gov/pubmed/23341825 pmid:23341825
- 12. Moritz S, Birkner C, Kloss M, Jahn H, Hand I, Haasen C, et al. Executive functioning in obsessive-compulsive disorder, unipolar depression, and schizophrenia. Arch Clin Neuropsychol [Internet]. Oxford University Press; 2002 Jul 1 [cited 2018 Nov 15];17(5):477–83. Available from: https://academic.oup.com/acn/article-lookup/doi/10.1093/arclin/17.5.477 pmid:14592001
- 13. Kennedy KJ. Age effects on Trail Making Test performance. [Internet]. Vol. 52, Perceptual and motor skills. 1981 [cited 2018 Sep 10]. Available from: http://www.ncbi.nlm.nih.gov/pubmed/7255077
- 14. Salthouse TA, Toth J, Daniels K, Parks C, Pak R, Wolbrette M, et al. Effects of Aging on Efficiency of Task Switching in a Variant of the Trail Making Test. Neuropsychology [Internet]. 2000 [cited 2018 Sep 10];14(1):102–11. Available from: https://s3.amazonaws.com/academia.edu.documents/30699292/Salthouse-Connections-N-2000.pdf?AWSAccessKeyId=AKIAIWOWYYGZ2Y53UL3A&Expires=1536613557&Signature=YWKpJHrgpOc7VBHL5aXpvWvxSaY%3D&response-content-disposition=inline%3B filename%3DEffects_of_aging_on_efficiency_of_task_s.pdf pmid:10674802
- 15. Drane DL, Yuspeh RL, Huthwaite JS, Klingler LK. Demographic characteristics and normative observations for derived-trail making test indices. Neuropsychiatry Neuropsychol Behav Neurol [Internet]. 2002 Mar [cited 2018 Nov 15];15(1):39–43. Available from: http://www.ncbi.nlm.nih.gov/pubmed/11877550 pmid:11877550
- 16. Tombaugh TN. Trail Making Test A and B: Normative data stratified by age and education. Arch Clin Neuropsychol [Internet]. No longer published by Elsevier; 2004 Mar 1 [cited 2018 Jul 19];19(2):203–14. Available from: https://www.sciencedirect.com/science/article/pii/S0887617703000398 pmid:15010086
- 17. Ogawa S, Lee TM, Kay AR, Tank DW. Brain magnetic resonance imaging with contrast dependent on blood oxygenation. Proc Natl Acad Sci U S A [Internet]. National Academy of Sciences; 1990 Dec 1 [cited 2018 Sep 14];87(24):9868–72. Available from: http://www.ncbi.nlm.nih.gov/pubmed/2124706 pmid:2124706
- 18. Moll J, Oliveira-Souza R de, Moll FT, Bramati IE, Andreiuolo PA. The cerebral correlates of set-shifting: an fMRI study of the trail making test. Arq Neuropsiquiatr [Internet]. Associação Arquivos de Neuro-Psiquiatria; 2002 Dec [cited 2018 Jul 9];60(4):900–5. Available from: http://www.scielo.br/scielo.php?script=sci_arttext&pid=S0004-282X2002000600002&lng=en&tlng=en pmid:12563376
- 19. Jacobson SC, Blanchard M, Connolly CC, Cannon M, Garavan H. An fMRI investigation of a novel analogue to the Trail-Making Test. Brain Cogn [Internet]. 2011 [cited 2018 Jul 9];77(1):60–70. Available from: https://ac-els-cdn-com.myaccess.library.utoronto.ca/S0278262611001060/1-s2.0-S0278262611001060-main.pdf?_tid=8a7d69ef-0c2c-4abf-8007-5fd616ccd231&acdnat=1531172778_4da39206c8b62ddf97ff6db90c8d2f5d pmid:21782309
- 20. Allen MD, Owens TE, Fong AK, Richards DR. A functional neuroimaging analysis of the Trail Making Test-B: Implications for clinical application. Behav Neurol [Internet]. 2011 [cited 2018 Nov 15];24(2):159–71. Available from: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5377980/pdf/BN-2011-24-2-476893.pdf pmid:21606577
- 21. Zakzanis KK, Mraz R, Graham SJ. An fMRI study of the Trail Making Test. Neuropsychologia [Internet]. 2005 [cited 2018 Jul 9];43:1878–86. Available from: https://ac-els-cdn-com.myaccess.library.utoronto.ca/S0028393205001442/1-s2.0-S0028393205001442-main.pdf?_tid=c4a5fbcc-0451-46f7-983e-cdee125f3815&acdnat=1531172823_8425e0746f625088ef8b410c4e30a7b9 pmid:16168730
- 22. Tam F, Churchill NW, Strother SC, Graham SJ. A new tablet for writing and drawing during functional MRI. Hum Brain Mapp. Wiley Subscription Services, Inc., A Wiley Company; 2012 Feb 1;33(7):1750–1.
- 23. Churchill NW, Spring R, Afshin-Pour B, Dong F, Strother SC. An Automated, Adaptive Framework for Optimizing Preprocessing Pipelines in Task-Based Functional MRI. Bodurka J, editor. PLoS One [Internet]. Public Library of Science; 2015 Jul 10 [cited 2018 Dec 10];10(7):e0131520. Available from: pmid:26161667
- 24. Karimpoor M, Tam F, Strother SC, Fischer CE, Schweizer TA, Graham SJ. A computerized tablet with visual feedback of hand position for functional magnetic resonance imaging. Front Hum Neurosci. Frontiers; 2015 Mar 25;9:150. pmid:25859201
- 25. Karimpoor M, Churchill NW, Tam F, Fischer CE, Schweizer TA, Graham SJ. Tablet-Based Functional MRI of the Trail Making Test: Effect of Tablet Interaction Mode. Front Hum Neurosci. Frontiers; 2017 Oct 24;11:496. pmid:29114212
- 26. Shibuya-Tayoshi S, Sumitani S, Kikuchi K, Tanaka T, Tayoshi S, Ueno SI, et al. Activation of the prefrontal cortex during the Trail-Making Test detected with multichannel near-infrared spectroscopy. Psychiatry Clin Neurosci [Internet]. Wiley/Blackwell (10.1111); 2007 Dec 7 [cited 2018 Jul 9];61(6):616–21. Available from: pmid:18081621
- 27. Cabeza R, Anderson ND, Locantore JK, McIntosh AR. Aging gracefully: Compensatory brain activity in high-performing older adults. Neuroimage. 2002;17(3):1394–402. pmid:12414279
- 28. Grady CL, Springer M V., Hongwanishkul D, McIntosh AR, Winocur G. Age-related changes in brain activity across the adult lifespan. Vol. 18, Journal of Cognitive Neuroscience. 2006.
- 29. Oldfield RCC. The assessment and analysis of handedness: The Edinburgh inventory [Internet]. Vol. 9, Neuropsychologia. Pergamon Press; 1971 Mar [cited 2018 Sep 14]. Available from: https://www.sciencedirect.com/science/article/pii/0028393271900674
- 30. Nasreddine ZS, Phillips NA, Bedirian V, Charbonneau S, Whitehead V, Collin I, et al. The Montreal Cognitive Assessment, MoCA: a Brief Screening Tool for Mild Cognitive Impairment [Internet]. Vol. 53, J Am Geriatr Soc. 2005 Apr [cited 2018 Dec 10]. Available from: papers2://publication/uuid/F8CBA89F-EA44-44BC-89AD-B7F166D683B2
- 31. Cox RW. AFNI: Software for Analysis and Visualization of Functional Magnetic Resonance Neuroimages. Comput Biomed Res. 1996;29:162–73. pmid:8812068
- 32. Smith SM, Jenkinson M, Woolrich MW, Beckmann CF, Behrens TEJ, Johansen-Berg H, et al. Advances in functional and structural MR image analysis and implementation as FSL. Neuroimage [Internet]. 2004 Jan [cited 2018 Mar 5];23:S208–19. Available from: http://www.ncbi.nlm.nih.gov/pubmed/15501092 pmid:15501092
- 33. Campbell KL, Grigg O, Saverino C, Churchill N, Grady CL. Age differences in the intrinsic functional connectivity of default network subsystems. Front Aging Neurosci [Internet]. Frontiers Media SA; 2013 [cited 2018 Jun 18];5:73. Available from: http://www.ncbi.nlm.nih.gov/pubmed/24294203 pmid:24294203
- 34. Churchill NW, Strother SC. PHYCAA+: An optimized, adaptive procedure for measuring and controlling physiological noise in BOLD fMRI. Neuroimage [Internet]. 2013 Nov 15 [cited 2018 Mar 5];82:306–25. Available from: http://www.ncbi.nlm.nih.gov/pubmed/23727534 pmid:23727534
- 35. Mazziotta JC, Toga AW, Evans A, Fox P, Lancaster J. A probabilistic atlas of the human brain: theory and rationale for its development. The International Consortium for Brain Mapping (ICBM). Neuroimage. 1995 Jun;2(2):89–101. pmid:9343592
- 36. Strother SC, Anderson J, Hansen LK, Kjems U, Kustra R, Sidtis J, et al. The quantitative evaluation of functional neuroimaging experiments: The NPAIRS data analysis framework. Neuroimage [Internet]. 2002 [cited 2018 Jun 18];15(4):747–71. Available from: http://www.lacontelab.org/papers/npairs-data-analysis-framework.pdf pmid:11906218
- 37. Lindquist MA, Meng Loh J, Atlas LY, Wager TD. Modeling the hemodynamic response function in fMRI: efficiency, bias and mis-modeling. Neuroimage. 2009 Mar;45(1 Suppl):S187–98. pmid:19084070
- 38. Hagen K, Ehlis A-C, Haeussinger FB, Heinzel S, Dresler T, Mueller LD, et al. Activation during the Trail Making Test measured with functional near-infrared spectroscopy in healthy elderly subjects ☆. 2014 [cited 2018 Nov 14]; Available from: http://dx.doi.org/10.1016/j.neuroimage.2013.09.014
- 39. Müller LD, Guhn A, Zeller JBM, Biehl SC, Dresler T, Hahn T, et al. Neural correlates of a standardized version of the trail making test in young and elderly adults: A functional near-infrared spectroscopy study. Neuropsychologia [Internet]. 2014 [cited 2018 Jul 9];56(1):271–9. Available from: http://dx.doi.org/10.1016/j.neuropsychologia.2014.01.019
- 40. Yochim B, Baldo J, Nelson A, Delis DC. D-KEFS Trail Making Test performance in patients with lateral prefrontal cortex lesions. Artic J Int Neuropsychol Soc [Internet]. 2007 [cited 2018 Jul 9]; Available from: https://www.researchgate.net/publication/6312286
- 41. Bastos-Leite AJ, van der Flier WM, van Straaten ECW, Staekenborg SS, Scheltens P, Barkhof F. The Contribution of Medial Temporal Lobe Atrophy and Vascular Pathology to Cognitive Impairment in Vascular Dementia. Stroke [Internet]. 2007 Dec [cited 2018 Nov 15];38(12):3182–5. Available from: http://www.ncbi.nlm.nih.gov/pubmed/17962598 pmid:17962598
- 42. Oosterman JM, Vogels RLC, Van Harten B, Gouw AA, Poggesi A, Scheltens P, et al. Assessing mental flexibility: Neuroanatomical and neuropsychological correlates of the trail making test in elderly people. Clin Neuropsychol [Internet]. 2010 [cited 2018 Sep 10];24(2):203–19. Available from: http://www.tandfonline.com/action/journalInformation?journalCode=ntcn20 pmid:20162494
- 43. Oosterman JM, Vogels RLC, van Harten B, Gouw AA, Scheltens P, Poggesi A, et al. The role of white matter hyperintensities and medial temporal lobe atrophy in age-related executive dysfunctioning. Brain Cogn [Internet]. 2008 Nov [cited 2018 Nov 15];68(2):128–33. Available from: http://www.ncbi.nlm.nih.gov/pubmed/18450353 pmid:18450353
- 44. Leonards U, Sunaert S, Van Hecke P, Orban GA. Attention Mechanisms in Visual Search—An fMRI Study. J Cogn Neurosci [Internet]. 2000 Nov [cited 2018 Nov 15];12(supplement 2):61–75. Available from: http://www.ncbi.nlm.nih.gov/pubmed/11506648
- 45. Nobre AC, Coull JT, Walsh V, Frith CD. Brain activations during visual search: contributions of search efficiency versus feature binding. Neuroimage [Internet]. 2003 Jan [cited 2018 Nov 16];18(1):91–103. Available from: http://www.ncbi.nlm.nih.gov/pubmed/12507447 pmid:12507447
- 46. Culham JC, Cavina-Pratesi C, Singhal A. The role of parietal cortex in visuomotor control: What have we learned from neuroimaging? Neuropsychologia [Internet]. 2006 Jan [cited 2018 Nov 16];44(13):2668–84. Available from: http://www.ncbi.nlm.nih.gov/pubmed/16337974 pmid:16337974
- 47. Husain M, Nachev P. Space and the parietal cortex. Trends Cogn Sci [Internet]. Elsevier; 2007 Jan 1 [cited 2018 Nov 16];11(1):30–6. Available from: http://www.ncbi.nlm.nih.gov/pubmed/17134935 pmid:17134935
- 48. Fransson P, Marrelec G. The precuneus/posterior cingulate cortex plays a pivotal role in the default mode network: Evidence from a partial correlation network analysis. Neuroimage [Internet]. Academic Press; 2008 Sep 1 [cited 2018 Oct 5];42(3):1178–84. Available from: https://www.sciencedirect.com/science/article/pii/S1053811908007283 pmid:18598773
- 49. Greicius MD, Krasnow B, Reiss AL, Menon V. Functional connectivity in the resting brain: a network analysis of the default mode hypothesis. Proc Natl Acad Sci U S A [Internet]. National Academy of Sciences; 2003 Jan 7 [cited 2018 Oct 5];100(1):253–8. Available from: http://www.ncbi.nlm.nih.gov/pubmed/12506194 pmid:12506194
- 50. Greicius MD, Supekar K, Menon V, Dougherty RF. Resting-State Functional Connectivity Reflects Structural Connectivity in the Default Mode Network. Cereb Cortex [Internet]. Oxford University Press; 2009 Jan 1 [cited 2018 Oct 5];19(1):72–8. Available from: https://academic.oup.com/cercor/article-lookup/doi/10.1093/cercor/bhn059 pmid:18403396
- 51. Mason MF, Norton MI, Van Horn JD, Wegner DM, Grafton ST, Macrae CN. Wandering Minds: The Default Network and Stimulus-Independent Thought. Science (80-) [Internet]. 2007 Jan 19 [cited 2018 Oct 5];315(5810):393–5. Available from: http://www.ncbi.nlm.nih.gov/pubmed/17234951
- 52. Ernst LH, Schneider S, Ehlis AC, Fallgatter AJ. Functional near infrared spectroscopy in psychiatry: A critical review. J Near Infrared Spectrosc [Internet]. SAGE PublicationsSage UK: London, England; 2012 Feb 1 [cited 2018 Nov 15];20(1):93–105. Available from: http://journals.sagepub.com/doi/10.1255/jnirs.970
- 53. Churchill NW, Reddy Raamana P, Spring R, Strother SC. Optimizing fMRI preprocessing pipelines for block-design tasks as a function of age. 2017 [cited 2018 Dec 7]; Available from: http://dx.doi.org/10.1016/j.neuroimage.2017.02.028
- 54. Salthouse TA. When does age-related cognitive decline begin? Vol. 30, Neurobiology of Aging. Elsevier Inc.; 2009. p. 507–14. pmid:19231028
- 55. Lenehan ME, Summers MJ, Saunders NL, Summers JJ, Vickers JC. Relationship between education and age-related cognitive decline: A review of recent research. Psychogeriatrics. 2015 Jun;15(2):154–62. pmid:25516261
- 56. Wilson RS, Hebert LE, Scherr PA, Barnes LL, De Leon CFM, Evans DA. Educational attainment and cognitive decline in old age. Neurology. Lippincott Williams and Wilkins; 2009 Feb 3;72(5):460–5. pmid:19188578
- 57. Deng I, Chung L, Talwar N, Tam F, Churchill N, Schweizer TA, et al. Functional MRI of Letter Cancellation Task Performance in Older Adults. Front Hum Neurosci. Frontiers; 2019;13:97.
- 58. Talwar NA, Churchill NW, Hird MA, Pshonyak I, Tam F, Fischer CE, et al. The Neural Correlates of the Clock-Drawing Test in Healthy Aging. Front Hum Neurosci. Frontiers; 2019 Feb 5;13:25. pmid:30804769
- 59. Elias MF, Dore GA, Goodell AL, Davey A, Zilioli MKC, Brennan S, et al. Normative data for elderly adults: The maine-syracuse study. Exp Aging Res [Internet]. 2011 [cited 2018 Mar 8];37(2):142–78. Available from: https://www-tandfonline-com.myaccess.library.utoronto.ca/doi/pdf/10.1080/0361073X.2011.554511?needAccess=true pmid:21424955
- 60. Grady CL, Ma Maisog J, Horwitz B, Ungerleider LG, Mentis MJ, Salerno JA, et al. Age-related Changes in Cortical Blood Flow Activation during Visual Processing of Faces and Location Largay for the behavioral testing; Dr. Margaret Daube-Witherspoon for keeping the tomograph and automatic blood sampler working. J Neurosci [Internet]. 1994 [cited 2018 Mar 8];14(3):1450–62. Available from: http://www.jneurosci.org/content/jneuro/14/3/1450.full.pdf
- 61. Harada CN, Natelson Love MC, Triebel KL. Normal cognitive aging. Clin Geriatr Med [Internet]. NIH Public Access; 2013 Nov [cited 2018 Mar 8];29(4):737–52. Available from: http://www.ncbi.nlm.nih.gov/pubmed/24094294 pmid:24094294
- 62. Miller IN, Himali JJ, Beiser AS, Murabito JM, Seshadri S, Wolf PA, et al. Normative Data for the Cognitively Intact Oldest-Old: The Framingham Heart Study. Exp Aging Res [Internet]. NIH Public Access; 2015 [cited 2018 Mar 8];41(4):386–409. Available from: http://www.ncbi.nlm.nih.gov/pubmed/26214098 pmid:26214098
- 63. Dennis NA, Cabeza R. Neuroimaging of Healthy Cognitive Aging. Handb Aging Cogn [Internet]. [cited 2018 Mar 8]; Available from: https://www.taylorfrancis.com/books/9780203837665
- 64. Fjell AM, Walhovd KB. Structural Brain Changes in Aging: Courses, Causes and Cognitive Consequences. Rev Neurosci. 2010;21:187–221. pmid:20879692
- 65. Raz N, Lindenberger U, Rodrigue KM, Kennedy KM, Head D, Williamson A, et al. Regional Brain Changes in Aging Healthy Adults: General Trends, Individual Differences and Modifiers. Cereb Cortex [Internet]. Oxford University Press; 2005 Nov 1 [cited 2018 Mar 8];15(11):1676–89. Available from: http://academic.oup.com/cercor/article/15/11/1676/296890/Regional-Brain-Changes-in-Aging-Healthy-Adults pmid:15703252
- 66. Resnick SM, Pham DL, Kraut MA, Zonderman AB, Davatzikos C. Longitudinal Magnetic Resonance Imaging Studies of Older Adults: A Shrinking Brain. J Neurosci [Internet]. 2003 [cited 2018 Mar 8];23(8):3295–301. Available from: http://www.jneurosci.org/content/23/8/3295.long pmid:12716936
- 67. Peters R. Ageing and the brain. Postgrad Med J [Internet]. BMJ Publishing Group; 2006 Feb [cited 2018 Nov 16];82(964):84–8. Available from: http://www.ncbi.nlm.nih.gov/pubmed/16461469 pmid:16461469
- 68. Tam CWC, Burton EJ, McKeith IG, Burn DJ, O’Brien JT. Temporal lobe atrophy on MRI in Parkinson disease with dementia: a comparison with Alzheimer disease and dementia with Lewy bodies. Neurology [Internet]. Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology; 2005 Mar 8 [cited 2018 Nov 16];64(5):861–5. Available from: http://www.ncbi.nlm.nih.gov/pubmed/15753423 pmid:15753423
- 69. Visser PJ, Scheltens P, Verhey F. Medial temporal lobe atrophy and memory dysfunction as predictors for dementia in subjects with mild cognitive impairment Clinical rules View project. Artic J Neurol [Internet]. 1999 [cited 2018 Nov 16];246:477–85. Available from: https://www.researchgate.net/publication/225676569
- 70. Chan D, Fox NC, Scahill RI, Crum WR, Whitwell JL, Leschziner G, et al. Patterns of temporal lobe atrophy in semantic dementia and Alzheimer’s disease. Ann Neurol [Internet]. John Wiley & Sons, Ltd; 2001 Apr 1 [cited 2018 Nov 16];49(4):433–42. Available from: http://doi.wiley.com/10.1002/ana.92 pmid:11310620
- 71. Burton EJ, Barber R, Mukaetova-Ladinska EB, Robson J, Perry RH, Jaros E, et al. Medial temporal lobe atrophy on MRI differentiates Alzheimer’s disease from dementia with Lewy bodies and vascular cognitive impairment: a prospective study with pathological verification of diagnosis. Brain [Internet]. Oxford University Press; 2009 Jan 1 [cited 2018 Nov 16];132(1):195–203. Available from: https://academic.oup.com/brain/article-lookup/doi/10.1093/brain/awn298
- 72. Dickerson BC, Salat DH, Bates JF, Atiya M, Killiany RJ, Greve DN, et al. Medial temporal lobe function and structure in mild cognitive impairment. Ann Neurol [Internet]. John Wiley & Sons, Ltd; 2004 Jul 1 [cited 2018 Jan 22];56(1):27–35. Available from: pmid:15236399
- 73. Korf ESC, Wahlund L-O, Visser PJ, Scheltens P. Medial temporal lobe atrophy on MRI predicts dementia in patients with mild cognitive impairment. Neurology [Internet]. American Academy of Neurology; 2004 Jul 13 [cited 2018 Jan 22];63(1):94–100. Available from: http://www.ncbi.nlm.nih.gov/pubmed/15249617 pmid:15249617
- 74. Bellebaum C, Daum I. Cerebellar involvement in executive control. The Cerebellum [Internet]. 2007 [cited 2018 Nov 16];6(3):184–92. Available from: http://www.ncbi.nlm.nih.gov/pubmed/17786814 pmid:17786814
- 75. Gottwald B, Wilde B, Mihajlovic Z, Mehdorn HM. Evidence for distinct cognitive deficits after focal cerebellar lesions. J Neurol Neurosurg Psychiatry [Internet]. 2004 Nov 1 [cited 2018 Nov 16];75(11):1524–31. Available from: http://www.ncbi.nlm.nih.gov/pubmed/15489381 pmid:15489381
- 76. Bolceková E, Mojzeš M, Van Tran Q, Kukal J, Ostrý S, Kulišťák P, et al. Cognitive impairment in cerebellar lesions: a logit model based on neuropsychological testing. Cerebellum & ataxias [Internet]. BioMed Central; 2017 [cited 2018 Nov 16];4:13. Available from: http://www.ncbi.nlm.nih.gov/pubmed/28775852
- 77. Charlot V, Tzourio N, Zilbovicius M, Mazoyer B, Denis M. Different mental imagery abilities result in different regional cerebral blood flow activation patterns during cognitive tasks. Neuropsychologia. 1992 Jun;30(6):565–80. pmid:1641120
- 78. Grabner RH, Neubauer AC, Stern E. Superior performance and neural efficiency: The impact of intelligence and expertise. Brain Res Bull [Internet]. 2006 [cited 2018 Oct 5];69:422–39. Available from: https://www.ethz.ch/content/dam/ethz/special-interest/gess/ifv/professur-lehr-und-lernforschung/publikationen-stern/ab-2005/Artikel/grabner_neubauer_stern_2006.pdf pmid:16624674
- 79. Rypma B, Berger JS, Genova HM, Rebbechi D, D’Esposito M. Dissociating age-related changes in cognitive strategy and neural efficiency using event-related fMRI. Cortex [Internet]. 2005 [cited 2018 Oct 5];41(4):582–94. Available from: http://www.ncbi.nlm.nih.gov/pubmed/16042034 pmid:16042034
- 80. D’Esposito M, Zarahn E, Arguirre GK, Rypma B. The effect of normal ageing on the coupling of neural activity to the BOLD hemodynamic response. Neuroimage. 1999;10:6–14. pmid:10385577
- 81. Wright ME, Wise RG. Can Blood Oxygenation Level Dependent Functional Magnetic Resonance Imaging Be Used Accurately to Compare Older and Younger Populations? A Mini Literature Review. Front Aging Neurosci [Internet]. Frontiers; 2018 Nov 13 [cited 2019 Jul 22];10:371. Available from: https://www.frontiersin.org/article/10.3389/fnagi.2018.00371/full pmid:30483117