Beyond a Differential Diagnosis: Cognitive and Morphometric Decoding of Information Processing Speed in Senior Adults with DSM-5 Mild Neurocognitive Disorders

Abstract

Background: Processing speed has been highlighted as a diagnostic item for neurocognitive disorders (NCD) in DSM-5. The utility of information processing speed (IPS) enclosed with multiscale constructs in the diagnosis of NCD warrants exploration.

Objective: We aimed to investigate the IPS with two types of measurements in the patients with NCD due to vascular disease (NCD-vascular) and NCD due to Alzheimer’s disease (NCD-AD), and examine the associations between IPS measures and morphometric features.

Methods: The IPS was evaluated using trail making test (TMT) and flanker test (n = 204). Direct scores, derived scores, and reaction time (RT) were used as IPS measures. Further, surface-based morphometry cortical volume was calculated in a subsample (n = 44) with structural MRI data.

Results: All IPS measures showed a significant value to differentiate NCD patients from healthy subjects. Only mean RT could distinguish NCD-AD from NCD-vascular groups. TMT-B score and difference score were correlated with gray matter volume (GMV) of inferior frontal gyrus, precuneus and superior temporal cortex. Mean RT was associated with the GMV of post-central gyrus (r = –0.327, p = 0.035), and executive speed was associated with inferior frontal cortex (r = –0.475, p = 0.001), cingulate gyrus (r = –0.497, p = 0.001), and superior temporal gyrus (r = –0.36, p = 0.019).

Conclusion: The cognitive and morphometric correlates of IPS measures indicate that complex IPS might be decomposed into the domain-specific components with corresponding neural underpinnings. Our findings may also provide essential insights into the diagnostic item of NCD.

Keywords

Gray matter volume information processing speed morphometry neurocognitive disorders reaction time trail making test

INTRODUCTION

Processing speed is a well-known indicator of brain efficacy and may affect the quality of life in the individuals with neurodegenerative disease [1]. In theDiagnosis and Statistical Manual, 5th edition (DSM-5), processing speed, serving as a diagnostic item of neurocognitive disorders (NCD) [2], has been highlighted in the updated criteria at the first time. With the advances of assessment techniques, information processing speed (IPS) is conceptualized as a composite dynamic construct that accounts for different measurements and multiscale indices derived from paper-based and computerized tests [3]. Acrossprevious studies, slowed IPS is asserted as one of the primary cognitive deficits in older adults with mild cognitive impairment (MCI) [4]. Notably, it appears that the IPS measured by computerized tests seems more sensitive to cognitive decline in senior adults [5, 6], particularly the ones with preclinical dementia [1 , 8].

Studies focusing on cognitive aging, MCI, and dementia have provided extensive and robust evidence of the clinical application of IPS in diagnosing preclinical dementia [9]. However, the performance of IPS derived from either modality-specific or task-specific measurements may lead to heterogeneous results, which should not be interpreted in isolation. In spite of the evolved criteria, an emerging challenge for clinicians is to decode the IPS for a better understanding of its nature and a practical utility in the diagnosis and differential diagnosis of mild NCD.

Indeed, the approach to investigating IPS is constrained by the methods available for study. The methods for decoding IPS could be organized according to the types of mechanistic insights that each technique provides. Two common conceptualizations of IPS are pronounced in aging neuroscience particularly [10]: the first is from the level of cognitive function, and the second is from the level of brain morphometry.

Cumulatively, it is still unclear whether the IPS with different measurement scales can be used to differentiate the ones with different cognitive status and to what extent these measures correlate with brain morphometry. Thus, in current study we would detect the multifaceted features of complex IPS in major subtypes of NCD patients. Our primary aim was to examine the IPS performance by using paper-based trail making test (TMT) and computer-based flanker test. A secondary aim was to explore the associations between IPS measures and brain morphometry.

MATERIALS AND METHODS

Experiment 1: Cognitive decoding of IPS

Participants and methods

In Experiment 1, we recruited 204 community-dwelling adults aged from 65–80 years from another cohort study aiming to establish a detailed neurocognitive profile of Chinese senior adults [11]. The eligible participants were scheduled for a 1.5-h interview at a research center for cognition in Hong Kong. A structured neuropsychological battery was used to evaluate the global cognition and the major domains of neurocognitive function [12]. The Cantonese version of Mini-Mental State Examination (CMMSE), Montreal Cognitive Assessment Hong Kong version (HK MoCA), Alzheimer’s Disease Assessment Scale-Cognitive subscale (ADAS-Cog), and Clinical dementia rating scale (CDR) were employed to evaluate the global cognitive function.

Domains of neurocognitive function included [13]: (1) Complex attention: digit span forward (DSF) and trail making test part A (TMT-A); (2) Language: Chinese verbal fluency test (CVFT); (3) Perceptual-motor function: the subjects performed the commands as “tap each shoulder twice with two fingers keeping your eyes shut”; (4) Learning and memory: digit span backward (DSB) and delayed recall of words; (5) executive function: Chinese Trail making test part B (TMT-B); (6) Conceptualization: recognition of the abstraction with similarity and difference.

Cerebrovascular risks were evaluated by the Cumulative illness rating scale for geriatrics (CIRS) for the presence and severity of heart diseases, hypertension, hyperlipidemia, diabetes mellitus, atrial fibrillation, and anemia [14]. The Pittsburgh sleep quality index [15], Cornell scale for depression in dementia [16], and activity of daily living scale [17] were used to assess the sleeps disorders, depressive symptoms, and everyday functioning separately. All the measurements were conducted with Chinese instructions.

Criteria for selection of healthy and NCD groups are as follows: (1) Healthy controls (HC) refers to the adults with cognitive performance within 1.5 standard deviation (SD) of age-normal reference derived from the cohort study [18], which presented with CMMSE score greater than 28 and CDR score equal to 0. (2) The mild NCD patients are defined by the following three criteria [2]: evidence of modest decline in one or more neurocognitive domains, which was set as ≥1.5 SD below the neurocognitive performance of HC; no interference with independence in everyday activities; and no comorbid major psychiatric disorders. NCD due to vascular disease (NCD-vascular) and NCD due to Alzheimer’s disease (NCD-AD) patients fulfilled the criteria of NCD. The definition of NCD-vascular patients required more than two chronic cerebrovascular risks and reduced functioning across the neurocognitive domains except for memory. NCD-AD patients demonstrated declined functioning in memory. (3) The exclusion criteria for this study included the following: clinical dementia, defined as cases with CMMSE score below the local cutoff for dementia of 18 and below for illiterate elderly, 20 and below for those with one to two years of education, and 22 and below for subjects with more than two years of education [19]; and cases with sleep disorders, depressive symptoms, or history of neurological and psychiatric disorders.

Ethics issues

This study was conducted according to the principles in the Declaration of Helsinki. Ethics approval was obtained from the Joint Chinese University of Hong Kong –New territories East Cluster Clinical Research Ethics Committee (Joint CUHK-NTECCREC). Written informed consent from all participants was obtained before the assessment.

IPS measures

The performance of IPS was evaluated by two cognitive tests: The first one is TMT, as a paper-based test, containing two subtests: (1) TMT part A (TMT-A) requires to draw a continuous line linking randomly distributed numbers in an ascending order as fast as possible (Fig. 1a: Arabic number; Fig. 1b: Chinese number). (2) TMT part B (TMT-B) requires to draw a continuous line linking a series of Arabic numbers and Chinese number (Fig. 1c) in ascending and alphabetical order as fast as possible. A time limit of 180 s is stipulated for this test. TMT-A is primarily addressed as a measure of visual attention [20]. TMT-B is used as a measure of task-switching and executive function [21].

Fig.1

Different measurements of IPS, including Trail making test (TMT) part A with numbers (a), TMT part A with Chinese characters (b), TMT part B (c), and flanker test (d).

Direct scores and derived scores of TMT are used to assess the different facets of IPS. Direct scores are equal to the completion time (in seconds) of TMT-A and TMT-B, reflecting the global performance of IPS. Derived scores are the extensive measures of TMT performance, including two indices: (1) Difference score (B – A): the subtraction of TMT-A score from TMT-B score is used as a measure of the executive IPS [22]; (2) Ratio score (B: A), the TMT-B score divided by TMT-A score is used as a measure of cognitive flexibility [23]. Higher direct and derived scores of TMT correspond to the worse performanceof IPS.

The second one is flanker test (with arrows), as a computer-based test, containing 288 trials for evaluating IPS with presence of reaction time (RT) in milliseconds (ms) [24]. In a given trial of flanker test, a cross-fixation presents in the middle of the computer screen for 400 to 1600 ms (randomized), subsequently replaced for 100 ms by the warning cues. The target, a central arrow appears above or below the cross-fixation and was surrounded by two arrows on each side. Three types of flanker, as neutral, congruent, and incongruent (Fig. 1d), implicate the increased levels of cognitive demand. At the beginning of the test, all participants were instructed to decide whether a central arrow points to left or right, and press a left button of the mouse if the central arrow was pointing to left, or right button if it was pointing to right.

Following the illustration, all participants were instructed to respond as fast as they could to the direction of the flanker by clicking left or right button. RT, as the completion time of a given trial, is collected for calculating IPS. Mean RT across all correct trials is employed as a measure of general performance of IPS. Similar to derived scores of TMT, the subtraction of RT of congruent from the RT of incongruent is employed as a measure of executive speed.

Experiment 2: Morphometric decoding of IPS

Participants and methods

A subsample of forty-four adults recruited from Experiment 1 was invited to participate the structural magnetic resonance imaging (sMRI) study. High-resolution T1-weighted MRI images were acquired in Prince of Wales Hospital with a 3.0 Tesla Philips Achieva MRI scanner (Philips Medical System, Best, The Netherlands). Scanning time for each participant was 8.5 min. The acquisition of sMRI scans were configured with the following parameters [25]: axial acquisition with a 256×256×192 matrix, thickness = 1 mm, no gap, field of view (FOV) = 230 mm, TR = 2070 ms, TE = 3.93 ms, flip angle = 15°. The sequence yielded high quality isotropic images with the voxel size of 1 mm×1 mm×1 mm.

Surface-based morphometry (SBM) analysis

Cortical morphometry, as gray matter volume (GMV), was calculated by BrainSuite 14.0 (http://brainsuite.org/). To calculate the regional GMV individually, we followed the standard pipeline with default parameters [26]. Each participant’s sMRI scans were processed with the following steps: (1) motion correction; (2) intensity normalization; (3) removal of non-brain voxels; (4) segmentation of gray matter, white matter, and cerebrospinal fluid; (5) tessellation of the gray matter/white matter (GM/WM) boundary and automated topology correction. At each step, the results were visually inspected and manual interventions were performed when required to correct topological defects.

Statistical analyses

Homogeneity of variance test was used to evaluate the equality of variances among healthy, NCD-AD and NCD-vascular groups. Group-wise differences were tested either with χ² test for category variable or one-way analysis of variance (ANOVA) for continuous variables. Tukey method was used to perform post hoc multiple comparisons as needed. As to the flanker test, the median values of RT across the trials were used as raw scores to avoid the influence of outliers. The effect sizes, measured by partial eta squared (η²), were calculated for the outcome of group differences in ANOVA. Pearson correlation coefficients were used to detect the relationships between IPS measures and cognitive performance. The operating characteristic curve (ROC) analysis was conducted to evaluate the values of IPS measures in differentiating the adults with different cognitive status.

As to the regional GMV, we followed the Cendes methods [27] to adjust the heterogeneity of head size and conducted a correction of individual variance of total intracranial volume (TIV) through the following formula: Corrected SV = (MBV×SV)/IBV. MBV refers to the mean brain volume in the group, SV is the regional volume of gray matter, and IBV is the individual TIV. The BrainSuite code(http://neuroimage.usc.edu/neuro/Resources/BST_SVReg_Utilities) embedded in MATLAB was employed to assess the group-wise differences of GMV with false discovery rate (FDR) correction [28]. Pearson correlation coefficients were used to examine the relationships between age, neurocognitive performance, and regional GMV. Significance levels were set at p value less than 0.05. Bonferroni corrections were addressed to reduce the chances of obtaining false-positive results when conducting multiple pairwise tests. The χ² test, ANOVA, Pearson correlation coefficient, and ROC analysis were performed by IBM SPSS 20.

RESULTS

Experiment 1

Demographics and neurocognitive performance

The basic participant demographics, including age, gender, and years of education were similar for the three groups examined. The NCD subgroups showed a decline of global cognition (Table 1). The NCD-AD group had the worst performance in the domain of short-term memory (measured by delayed recall). The NCD-vascular group had marked and extensive cardiovascular burden with the presence of higher scores on the CIRS (Cardiovascular risk factor: F = 3.28, p = 0.04; Heart disease: F = 9.11, p < 0.001; Lipid: F = 4.395, p = 0.014). The NCD-vascular group also presented poorer executive function (measured by CVFT) than the NCD-AD group.

Table 1

Demographics and neurocognitive features in healthy and NCD groups

	Healthy	NCD-AD	NCD-Vascular	F	p value
	(n = 137)	(n = 36)	(n = 31)	(χ²)
Age	71.45±3.99	73.35±4.91	73.46±5.27	3.098	0.056
Gender (F/M)	61/76	16/20	12/19	1.800	0.168
Education (years)	9.30±4.26	7.56±3.92	8.21±3.27	3.141	0.055
CSDD	0.49±1.81	0.18±0.72	0.26±0.85	0.590	0.555
PSQI	5.75±3.15	6.24±3.69	5.88±3.05	1.284	0.279
ADL	6.09±3.37	5.19±3.12	6.35±3.27	3.139	0.061
CDR-SOB	0.34±0.47	0.81±0.54	0.81±0.70	17.915	<0.001
CMMSE	28.63±1.17	27.22±1.96	26.74±1.73	30.869	<0.001
HK MoCA	27.23±1.85	24.11±3.21	24.00±2.77	44.020	<0.001
ADAS-Cog	4.91±2.08	9.42±2.49	8.98±2.63	82.615	<0.001
Digit span forward	7.62±1.13	7.17±0.94	7.00±1.44	5.118	0.007
CVFT	47.71±9.18	39.81±8.26	35.29±5.96	32.51	<0.001
Delayed recall	7.64±1.45	3.97±0.91	5.97±1.08	115.717	<0.001
Digit span backward	3.75±1.32	3.00±1.22	3.03±0.84	7.863	0.001
Perceptual-motor function	0.14±0.35	0.31±0.53	0.55±0.72	10.342	<0.001
Social cognition	1.52±0.54	1.39±0.59	1.35±0.601	1.6	0.204

Data are raw scores and presented as mean±SD. CSDD, The Cornell Scale for Depression in Dementia; PSQI, Pittsburgh Sleep Quality Index; CVFT, Chinese verbal fluency test; ADL, Activity of daily living scale; CDR-SOB, Clinical dementia rating-sum of box; CMMSE, Cantonese version of Mini-Mental State Examination; HK MoCA, Montreal Cognitive Assessment Hong Kong version; ADAS-Cog, Alzheimer’s Disease Assessment Scale-Cognitive subscale.

Evaluation of IPS

NCD patients had the worse performance on TMT and flanker test, including direct scores (TMT-A: HC: 12.89±6.37 s, NCD: 16.29±8.49 s, t = –2.89, p = 0.005; TMT-B: HC: 67.07±41.26 s, NCD: 87.44±52.11 s, t = –3.01, p = 0.003; Mean RT: HC: 687.49±105.97 ms, NCD: 720.39±121.18 ms, t =–1.99, p = 0.048) and derived scores (Difference score: HC: 54.22±38.19, NCD: 71.29±47.53 ms, t = –2.74, p = 0.007; Executive speed: HC: 59.86±45.86, NCD: 93.22±49.71, t = –4.75, p < 0.001).

As to the NCD subgroups, NCD-vascular and NCD-AD group presented poor performance on TMT and flanker test (Table 2). Post-hoc analyses found that both NCD-vascular and NCD-AD had prominent slowed IPS than HC. Meanwhile NCD-AD showed similar performance on TMT, but slowed mean RTs than NCD-vascular across the conditions with three flanker types.

Table 2

Comparisons of IPS measures between healthy and NCD groups

	Healthy	NCD-Vascular	NCD-AD	F	p value	Partial eta
	(n = 137)	(n = 31)	(n = 36)			squared (η²)
TMT
TMT-A (s)	11.39±5.54	14.23±6.03	15.67±10.91	6.610	0.002	0.05
TMT-B (s)	67.07±41.26	83.17±38.13	91.21±62.26	4.784	0.009	0.05
Difference score (B-A)	54.22±38.19	67.56±34.16	74.59±57.15	3.990	0.02	0.04
Ratio score (B: A)	5.54±3.37	5.55±2.26	5.92±2.25	0.223	0.801	0.002
Flanker test
RT of neutral	652.69±105.78	644.34±91.82	707.96±120.81	4.284	0.015	0.04
RT of congruent	666.39±109.8	653.42±92.29	715.01±135.85	3.218	0.042	0.03
RT of incongruent	726.26±107.88	742.29±108.84	811.96±143.06	7.93	<0.001	0.07
Mean RT (ms)	687.49±105.98	685.32±97.09	750.59±132.59	4.958	0.008	0.047
Executive speed	59.86±45.86	88.87±44.84	96.96±53.90	11.472	<0.001	0.102

Data are raw scores and presented as mean±SD. TMT-A, Trail making test part A; TMT-B, Trail making test part B.

Associations between age, IPS, and neurocognitive function

Within healthy adults, prominent age-related decline has been found in TMT-B score (r = 0.237, p = 0.005), difference score of TMT (r = 0.235, p = 0.005) and executive speed of flanker test (r = 0.257, p = 0.002). No significant age-related IPS change was found in NCD patients. Using age, gender, and years of education as covariates, prominent associations were found between IPS measures and neurocognitive performance in both healthy and NCD groups. As to TMT indices (Table 3), direct and difference scores were negatively correlated with the global cognition measured by CDR and HK MoCA. Instead, mean RTs under the conditions with three flanker types and executive speed derived from flanker test were negatively associated with the episodic memory measured by delayed recall (Table 4). Interestingly, the Pearson correlation coefficients between mean RTs and delayed recall score were increased in parallel with the level of cognitive demand (RT of Neutral: r = –0.2, p = 0.005, RT of Congruent: r = –0.217, p = 0.002, RT of Incongruent: r = –0.249, p < 0.001). Meanwhile, no significant relationship was found between TMT indices and flanker test indices, which indicting the IPS measures derived from TMT and flanker test may represent the different facets of IPS and relatively independent from each other.

Table 3

Correlations of TMT indices and neurocognitive features

	TMT-A		TMT-B		Difference score		Ratio score
	r	p	r	p	r	p	r	p
CDR-SOB	0.194	0.006	0.259	<0.001	0.245	<0.001	0.125	0.079
CMMSE	–0.105	0.141	–0.176	0.013	–0.172	0.015	–0.080	0.260
HK MoCA	–0.222	0.002	–0.302	<0.001	–0.287	<0.001	–0.093	0.191
ADAS-Cog	0.050	0.487	0.131	0.066	0.132	0.064	0.103	0.149
Delayed recall	–0.041	0.565	–0.099	0.165	–0.099	0.164	–0.053	0.459
Digit span backward	0.027	0.706	–0.119	0.094	–0.132	0.063	–0.171	0.016

Table 4

Correlations of flanker test indices and neurocognitive features

	RT of neutral		RT of congruent		RT of incongruent		Executive speed
	r	p	r	p	r	p	r	p
CDR-SOB	0.066	0.35	0.037	0.606	0.032	0.65	–0.007	0.918
CMMSE	0.034	0.632	0.03	0.669	–0.015	0.836	–0.107	0.133
HK MoCA	–0.094	0.187	–0.088	0.216	–0.121	0.087	–0.089	0.21
ADAS-Cog	0.08	0.26	0.059	0.403	0.124	0.08	0.138	0.052
Delayed recall	–0.2	0.005	–0.217	0.002	–0.249	<0.001	–0.169	0.016
Digit span backward	0.009	0.905	–0.008	0.908	–0.016	0.821	–0.02	0.78

ROC analysis for IPS measures

As Supplementary Figure 1 shown, IPS measures, including direct scores (TMT-A: AUC value = 0.625, p = 0.004, 95% confidence interval (CI): 0.541–0.708; TMT-B: AUC value = 0.661, p < 0.001, 95% CI: 0.583–0.739), difference score (AUC value = 0.663, p < 0.001, 95% CI: 0.586–0.741), and executive speed (AUC value = 0.689, p < 0.001,95% CI: 0.612–0.756), showed significant power to differentiate NCD patients from HC. As to NCD subgroups, IPS measures derived from TMT (TMT-B: AUC value = 0.652, p = 0.006, 95% CI: 0.553–0.751; difference score: AUC value = 0.653, p = 0.005, 95% CI: 0.556–0.75; ratio score: AUC value = 0.614, p = 0.038, 95% CI: 0.508–0.72) and flanker test (RT of congruent: AUC value = 0.639, p = 0.01, 95% CI: 0.537–0.742; RT of incongruent: AUC value = 0.68, p = 0.001; 95% CI: 0.58–0.78) showed a moderate power to discriminate NCD-AD from HC.

Between HC and NCD-vascular, the TMT indices (TMT-A: AUC value = 0.658, p = 0.006, 95% CI: 0.559–0.758; TMT-B: AUC value = 0.672, p = 0.003, 95% CI: 0.575–0.768; difference score: AUC value = 0.675, p = 0.002; 95% CI: 0.578–0.772) showed a moderate power. Likewise, the executive speed measured by flanker test also demonstrated a moderate value (AUC value = 0.662, p = 0.005; 95% CI: 0.56–0.763). Notably, between NCD-AD and NCD-vascular patients, none of TMT indices presented a significant value. Instead, mean RT (AUC value = 0.652, p = 0.034, 95% CI: 0.52–0.784) and RT of neutral (AUC value = 0.656, p = 0.029, 95% CI: 0.525–0.788) showed a slight but significant value in differentiating NCD-AD fromNCD-vascular.

Experiment 2

Demographics, neurocognitive features, and IPS

Consistent with the results of Experiment 1, NCD patients had worse performance on TMT and flanker test in terms of direct scores (TMT-B: HC: 51.11±27.93 s, NCD: 84.74±44.77 s, t = –3.053, p = 0.004; Mean RT: HC: 623.36±70.21 ms, NCD: 732.35±112.01 ms, t = –3.347, p = 0.004) and derived scores (difference score: HC: 39.84±23.34, NCD: 69.54±40.23 ms, t = –2.568, p = 0.02; executive speed: HC: 63.82±47.53, NCD: 101.25±70.03, t = –2.085, p = 0.04).

Comparisons of cortical morphometry

NCD patients demonstrated globally cortical atrophy, but with a marked reduction in GMV in the orbital frontal gyrus (HC: 2495.99±319.49 mm³, NCD: 2264.15±298.66 mm³, t = 2.287, p = 0.027), lingual gyrus (HC: 4755.23±615.76 mm³, NCD: 4082.68±644.63 mm³, t = 3.325, p = 0.002), superior temporal gyrus (HC: 9095.73±918.22 mm³, NCD: 7835.41±660.83 mm³, t = 4.597, p < 0.001) and temporal pole (HC: 5408.37±1059.99 mm³, NCD: 4462.39±1195.31 mm³, t = 2.648, p = 0.011).

Associations between age, IPS, and brain morphometry

Prominent age-related atrophy was found in temporal pole (r = –0.43, p = 0.004), paracentral lobule (r = –0.344, p = 0.024), and hippocampus (r = –0.451, p = 0.009). Using age, gender, years of education, and TIV as covariates, the correlations between IPS measures and regional GMV presented the domain-specific patterns. As Fig. 2 depicted, TMT-A score was correlated with left posterior orbitofrontal gyrus (POrG) (r = –0.334, p = 0.031); TMT-B score was correlated with left precuneus (r = –0.397, p = 0.009) and right lateral orbitofrontal gyrus (LOrG) (r = –0.35, p = 0.023). Derived scores of TMT were associated with right LOrG (difference score: r = –0.341, p = 0.027), left superior temporal gyrus (STG) (difference score: r = –0.345, p = 0.025) and left precuneus (difference score: r = –0.408, p = 0.007; ratio score: r = –0.321, p = 0.038).

Fig.2

Linear associations between regional gray matter volume (GMV) and IPS measures derived from TMT.

As Fig. 3 shown, IPS measured by flanker test was correlated with right post-central gyrus (PoG) (Mean RT: r = –0.327, p = 0.035), left inferior frontal gyrus (IFG) (executive speed: r = –0.475, p = 0.001), left posterior cingulate gyrus (PCC) (executive speed: r = –0.497, p = 0.001) and right STG (executive speed: r = –0.36, p = 0.019).

Fig.3

Linear associations between regional gray matter volume (GMV) and IPS measures derived from flanker test.

Considering the potential impacts of cardiovascular risks, the associations between CIRS score and brain morphometry were also conducted. Higher cardiovascular burden was negatively correlated with the GMV of left IFG (r = –0.408, p = 0.007). Thus, after adding CIRS score into covariates, the patterns of “IPS-GMV” correlates remained, however, the Pearson correlation coefficients between IPS measures and regional GMV became attenuated, including TMT-A and left POrG (r = –0.321, p = 0.041); TMT-B and right LOrG (r = –0.343, p = 0.028), left precuneus (r = –0.394, p = 0.011); difference score and right LOrG (r = –0.334, p = 0.033), left STG (r = –0.331, p = 0.034) and left precuneus (r = –0.404, p = 0.009); ratio score and left precuneus (r = –0.318, p = 0.043); executive speed and left IFG (r = –0.443, p = 0.004), left PCC (r = –0.459, p = 0.003) and right STG (r = –0.353, p = 0.024).

DISCUSSION

In this study, we conducted both traditional and computerized measurements of information processing speed in DSM-5 NCD patients and assessed their relationships with neurocognitive function and cortical morphometry. Overall, the majority of our findings are aligned with prior evidence that the ones with cognitive deficits showed worse performance on TMT [1 , 29] and flanker test [30, 31]. The disturbance of IPS in NCD-AD, compared with NCD-vascular, appears to demonstrate a seemingly different pattern with task-specific features. Moreover, the IPS measures derived from TMT and flanker test are complementary in differentiation of NCD-AD and NCD-vascular, which are also linked to the shared neuroanatomical underpinnings.

Complex IPS with cognitive scale

It is well accepted that the performance on IPS has deteriorated with advancing age, and correspondingly associated with other cognitive components [32, 33]. From our observations, given the age-related slowing in speed, two distinct patterns “IPS-cognition “correlates were detected in healthy and NCD groups. Worse global cognition was associated with slowed IPS assessed by TMT in cognitively normal senior adults; while in NCD patients, poorer working memory correlated with slowed IPS measured by flanker test. Indeed, of studies that have examined IPS and cognitive abilities, most reported a prominent relationship between declined working memory and slowed IPS in both young and older populations [34, 35]. However, some studies do not find a direct link between IPS and working memory [36, 37]. Based on the viewpoint from Salthouse [38], it is noteworthy to rethink that the determinants of IPS might contribute to the age-related cognitive decline. Even though no neuropsychological IPS measures are entirely pure; that is, they all tap other components, which may subserve other higher-order cognitive function [39]. Of note, in NCD patients, the observed link between mean RT and delayed recall score highlights the importance of IPS in working memory function, even in the absence of age-related memory decline. Compared to healthy controls, the different “IPS-cognition” correlation pattern in NCD patients might reflect the potential counterbalance of brain function under clinical state. Likewise, the heterogeneous associations between IPS indices and cognitive function also indicate that IPS, as a multifaceted construct, may be a core determinant of domain-specific ability in the individuals with different cognitive status.

Older adults with NCD-vascular demonstrated impaired executive facet of IPS, even the compensatory strategies at early stages may still allow this group to maintain a level of mean RT similar to that of healthy ones. In addition, the ones with NCD-AD have suffered from more breakdowns of IPS, including both general and executive speed across all IPS measures. Thus, it appears that the subtypes of NCD might possess shared and specific deficits on IPS. Indeed, the results from ROC analysis confirmed that the direct and difference scores of TMT, and the executive speed of flanker test had a moderate power to distinguish NCD patients from healthy cases. As to NCD subgroups, all IPS measures could differentiate NCD-AD patients from healthy controls, and executive speed derived from flanker test and TMT could discriminate NCD-vascular patients from the controls. Importantly, however, only mean RT of flanker test could distinguish NCD-AD from NCD-vascular, which is considered as one of the great difficulties in clinical practice [30].

Considering a growing consensus of differences in IPS partially due to impaired cognitive abilities [1 , 41], particularly, the abilities as episodic memory and executive function are the key determinants of the diagnostic criteria of DSM-5 NCD-AD and NCD-vascular. Thus, the observed results support the possibility to explain the differential roles of IPS with at least two cognitive components: one is general speed with the presence of TMT-A score and mean RT, another is executive speed with the presence of TMT-B score, difference score, and executive RT.

Complex IPS with morphometric scale

Similar to cognitive scale, the morphometric correlates of IPS also reflect the heterogeneous features across the IPS measures derived from the different neuropsychological tasks. The prior identified “IPS-brain” associations implied for a variety of age-related differences. For instance, some voxel-based morphometry studies found no correlation between gray matter and RT in middle-aged and older participants [42]. On the other hand, some studies reported that RT was strongly correlated with the cortical volume in frontal cortex, cingulate gyrus, and medial temporal lobe in adolescents [43] and young adults [44].

The aforementioned findings indicate that there is no single brain region that is invariably associated with IPS. Rather, the well-established brain mechanisms underlying specific cognitive function correspond to the regions related to the domain-specific IPS measures, particularly the executive speed. Regarding to the executive facet of IPS, disturbed executive speed (i.e., TMT-B) is accompanied with reduced cortical volume in frontal [45 –47] and temporal lobe [44] broadly, anterior cortical regions specifically. Based on the whole-brain analysis, our finding of interest is that the brain regions underlying the complex IPS are not necessarily overlapping with the given areas of age-related changes. Thus, the “IPS-GMV” correlations may not just reflect the potential anatomical underpinnings, but also reshape the domain-specific IPS indices with morphometric scale. The general IPS evaluated by different tasks presented relatively higher variability within correlation patterns with brain volume. Meanwhile, the cortical areas associated with executive speed were highly consistent across the IPS measures, which may extend prior work and provide support for the emphasis of “IPS-brain” correlations in abovementioned cognitive components of complex IPS.

IPS and NCD: Implications for clinicians

As exemplified by outlining the complex IPS with cognitive scale, the differentiation between disease-specific NCD has been achieved by means of paper-based and computer-based measurements. While considering NCD-AD and NCD-vascular encompassed multiple possible etiology with substantial overlap of cognitive impairments at early stage, the differential effects implemented in multiple IPS measures might require a more precise scale, such as millisecond.

Because of the linkages between domain-specific impairments and diagnostic classifications of NCD subgroups are not encoded in DSM-5 [48], it may be difficult to achieve a consensus on cut-off score and delineation of specific neurocognitive function [49], not even mentioned on the various neuropsychological tests with multiscale measures. It is intriguing to note that processing speed subsevers as a diagnostic item affiliated with complex attention; however, the associations between IPS indices and cognitive abilities in NCD patients were only found in working memory, of which the absolute values of correlation coefficients became greater with increased cognitive demand (i.e. from neutral to incongruent flanker). Just as our preceding results implicated, the components of IPS, as a cognitive capacity [50], might contribute to the attention [51 –53] and depth of processing [54] during episodic encoding (i.e., delayed recall).

Back to the clinical utility of complex IPS, the forthcoming challenge for collecting IPS in the classification of disease-specific NCD is in a transition stage of adapting evaluation to maintain comparable across different measurement scales. In sum, a combination of traditional and computerized assessments presents pros and cons to evaluate the multifaceted nature of IPS in healthy and clinical population. It seems that IPS indices derived from computerized task give a better fit to the cognitive and morphometric correlates, and a practical utility in differentiating the individuals with different cognitive status.

Limitations

Despite providing a link between multifaceted IPS and diagnostic classifications of DSM-5 NCD, the results of this study should be interpreted with its limitations. First, the sample size of each subgroup was unbalanced. We tested the homogeneity of variance before performing the ANOVA to detect the group-wise differences and address this issue. Second, the cross-sectional design had very limited power to infer any aging influence. Third, the classifications of healthy and NCD subgroups were based on neuropsychological battery, not ascertained by neuroimaging investigations. Additionally, the absence of MRI in the diagnosis of NCD-vascular group patients was a major limitation. We conducted detailed CIRS to infer the condition of cerebrovascular burden of each participant. Fourth, the cross-sectional design had limited power to infer any causative influence of the above findings. Future studies should involve high-resolution T2-weighted MRI, and perform a cross-validation of the diagnosis in NCD-vascular cases.

Footnotes

ACKNOWLEDGMENTS

This research was supported by the Lui Che Woo Institute of Innovation Medicine Grant at The Chinese University of Hong Kong. The authors thank Ms. Wing Yan Law and Ms. Ka Wun Chan from Department of Psychiatry and Professor Wang Defeng, Professor Chu Chiu Wing Winnie, and Dr. Lou Wutao from Department of Imaging and Intervention Radiology for their great efforts of collecting neurocognitive and neuroimaging data. The authors also thank all the participants and their families to participate this study. And we also thank the reviewers for their valuable comments and suggestions to improve the quality of this paper.

Authors’ disclosures available online ().

Appendix

References

Haworth

, Phillips

, Newson

, Rogers

, Torrens-Burton

, Tales

(2016) Measuring information processing speed in mild cognitive impairment: Clinical versus research dichotomy. J Alzheimers Dis 51, 263–275.

American Psychiatric Association (2013) Diagnostic and statistical manual of mental disorders (DSM-5^®). American Psychiatric Association, Washington, DC.

Salthouse

(1996) Constraints on theories of cognitive aging. Psychon Bull Rev 3, 287–299.

Hedden

, Gabrieli

(2004) Insights into the ageing mind: A view from cognitive neuroscience. Nat Rev Neurosci 5, 87–96.

Artero

, Petersen

, Touchon

, Ritchie

(2006) Revised criteria for mild cognitive impairment: Validation within a longitudinal population study. Dement Geriatr Cogn Disord 22, 465–470.

Barnes

, Yaffe

, Belfor

, Jagust

, DeCarli

, Reed

, Kramer

(2009) Computer-based cognitive training for mild cognitive impairment: Results from a pilot randomized, controlled trial. Alzheimer Dis Assoc Disord 23, 205–210.

Owsley

(2013) Visual processing speed. Vision Res 90, 52–56.

Gorus

, De Raedt

, Lambert

, Lemper

, Mets

(2008) Reaction times and performance variability in normal aging, mild cognitive impairment, and Alzheimer’s disease. J Geriatr Psychiatry Neurol 21, 204–218.

Wild

, Howieson

, Webbe

, Seelye

, Kaye

(2008) Status of computerized cognitive testing in aging: A systematic review. Alzheimer Dement 4, 428–437.

10.

Lockhart

, DeCarli

, Fama

(2014) Neuroimaging of the aging brain: Introduction to the secial issue of Neuropsychology Review. Neuropsychol Rev 24, 267–270.

11.

Lam

, Chan

, Leung

, Fung

, Leung

(2015) Would olderadults with mild cognitive impairment adhere to and benefit from astructured lifestyle activity intervention to enhance cognition? Acluster randomized controlled trial. PLoS One 10, e0118173.

12.

, Fung

, Chan

, Lam

(2016) Disturbance of attention network functions in Chinese healthy older adults: An intra-individual perspective. Int Psychogeriatr 28, 291–301.

13.

Sachdev

, Blacker

, Blazer

, Ganguli

, Jeste

, Paulsen

, Petersen

(2014) Classifying neurocognitive disorders: The DSM-5 approach. Nat Rev Neurosci 10, 634–642.

14.

Chan

, Chiu

, Lam

, Wong

, Conwell

(2014) A multidimensional risk factor model for suicide attempts in later life. Neuropsychiat Dis Treat 10, 1807–1817.

15.

Carpenter

, Andrykowski

(1998) Psychometric evaluation of the Pittsburgh sleep quality index. J Psychosom Res 45, 5–13.

16.

Schreiner

, Hayakawa

, Morimoto

, Kakuma

(2003) Screening for late life depression: Cut-off scores for the Geriatric Depression Scale and the Cornell Scale for Depression in Dementia among Japanese subjects. Int J Geriatr Psychiatry 18, 498–505.

17.

Lam

, Tam

, Chiu

, Lui

(2007) Depression and apathy affect functioning in community active subjects with questionable dementia and mild Alzheimer’s disease. Int J Geriatr Psychiatry 22, 431–437.

18.

, Chan

, Lam

(2016) Associations between intra-individual variability of reaction time and cognitive function in cognitively normal senior adults: Still beyond good or bad? Geriatrics 1, 13.

19.

Lam

, Tam

, Lui

, Chan

, Wong

, Tham

, Ho

, Chan

, Chiu

(2008) Prevalence ofvery mild and mild dementia in community-dwelling older Chinese people in Hong Kong. Int Psychogeriatr 20, 135–148.

20.

Kim

, Baek

, Kim

(2014) Alternative type of the trail making test in nonnative english-speakers: The trail making test-black & white. PLoS One 9, 1–6.

21.

Arbuthnott

, Frank

(2000) Trail making test, part B as a measure of executive control: Validation using a set-switching paradigm. J Clin Exp Neuropsychol 22, 518–528.

22.

Sanchez-Cubillo

, Perianez

, Adrover-Roig

, Rodriguez-Sanchez

, Rios-Lago

, Tirapu

, Barcelo

(2009) Construct validity of the Trail Making Test: Role of task-switching,working memory, inhibition/interference control, and visuomotor abilities. J Int Neuropsychol Soc 15, 438.

23.

Terada

, Sato

, Nagao

, Ikeda

, Shindo

, Hayashi

, Oshima

, Yokota

, Uchitomi

(2013) Trail Making Test B and brain perfusion imaging in mild cognitive impairment and mild Alzheimer’s disease. Psychiatry Res 213, 249–255.

24.

Eriksen

(1995) The flankers task and response competition: A useful tool for investigating a variety of cognitive problems. Vis Cogn 2, 101–118.

25.

, Ma

, Chan

, Lam

(2016) The effects of apolipoprotein ɛ 4 on aging brain in cognitively normal Chinese elderly: A surface-based morphometry study. Int Psychogeriatr 28, 1503–1511.

26.

Shattuck

, Mirza

, Adisetiyo

, Hojatkashani

, Salamon

, Narr

, Poldrack

, Bilder

, Toga

(2008) Construction of a 3D probabilistic atlas of human cortical structures. Neuroimage 39, 1064–1080.

27.

Cendes

, Andermann

, Gloor

, Evans

, Jones-Gotman

, Watson

, Melanson

, Olivier

, Peters

, Lopes-Cendes

, et al. (1993) MRIvolumetric measurement of amygdala and hippocampus in temporallobe epilepsy. Neurology 43, 719–719.

28.

Reiter

, Alpert

, Cobia

, Kwasny

, Morris

, Csernansky

, Wang

(2012) Cognitively normal individuals with AD parents may be at risk for developing aging-related cortical thinning patterns characteristic of AD. Neuroimage 61, 525–532.

29.

Ashendorf

, Jefferson

, O’Connor

, Chaisson

, Green

, Stern

(2008) Trail Making Test errors in normal aging, mild cognitive impairment, and dementia. Arch Clin Neuropsychol 23, 129–137.

30.

Fernáandez

, Campoy

, García Santos

, Antequera

, García-Sevilla

, Castillo

, Antúnez

, Fuentes

(2011) Is there a specific pattern of attention deficit in mild cognitive impairmentwith subcortical vascular features? Evidence from the Attention Network Test. Dement Geriatr Cogn Disord 31, 268–275.

31.

Van Dam

, Sano

, Mitsis

, Grossman

, Gu

, Park

, Hof

, Fan

(2013) Functional neural correlates of attentional deficitsin amnestic mild cognitive impairment. PLoS One 8, e54035.

32.

Cerella

, Hale

(1994) The rise and fall ininformation-processing rates over the life span. Acta Psychol(Amst) 86, 109–197.

33.

Chaytor

, Schmitter-Edgecombe

(2004) Working memory and aging: A cross-sectional and longitudinal analysis using a self-ordered pointing task. J Int Neuropsychol Soc 10, 489–503.

34.

Nebes

, Butters

, Mulsant

, Pollock

, Zmuda

, Houck

, Reynolds

(2000) Decreased working memory and processing speed mediate cognitive impairment in geriatric depression. Psychol Med 30, 679–691.

35.

Lengenfelder

, Bryant

, Diamond

, Kalmar

, Moore

, DeLuca

(2006) Processing speed interacts with working memory efficiency in multiple sclerosis. Arch Clin Neuropsychol 21, 229–238.

36.

Perlstein

, Cole

, Demery

, Seignourel

, Dixit

, Larson

, Briggs

(2004) Parametric manipulation of working memory load in traumatic brain injury: Behavioral and neural correlates. J Int Neuropsychol Soc 10, 724–741.

37.

Salmond

, Chatfield

, Menon

, Pickard

, Sahakian

(2005) Cognitive sequelae of head injury: Involvement of basal forebrain and associated structures. Brain 128, 189–200.

38.

Salthouse

(1994) The aging of working memory. Neuropsychology 8, 535.

39.

DeLuca

, Kalmar

(2013) Information processing speed in clinical populations, Psychology Press.

40.

Fernandez-Duque

, Black

(2006) Attentional networks in normal aging and Alzheimer’s disease. Neuropsychology 20, 133–143.

41.

Mahoney

, Verghese

, Goldin

, Lipton

, Holtzer

(2010) Alerting, orienting, and executive attention in older adults. J Int Neuropsychol Soc 16, 877–889.

42.

Haier

, Jung

, Yeo

, Head

, Alkire

(2005) Structural brain variation, age, and response time. Cogn Affect Behav Neurosci 5, 246–251.

43.

McAlonan

, Cheung

, Chua

, Oosterlaan

, Hung

, Tang

, Lee

, Kwong

, Ho

, Cheung

, Suckling

, Leung

(2009) Age-related grey matter volume correlates of response inhibition and shifting inattention-deficit hyperactivity disorder. Br J Psychiatry 194, 123–129.

44.

Chen

, Yang

, Lai

, Li

, Yuan

(2015) Correlating gray mattervolume with individual difference in the flanker interferenceeffect. PLoS One 10, e0136877.

45.

Smith

, Jonides

(1999) Storage and executive processes in the frontal lobes. Science 283, 1657–1661.

46.

Zimmerman

, Brickman

, Paul

, Grieve

, Tate

, Gunstad

, Cohen

, Aloia

, Williams

, Clark

, Whitford

, Gordon

(2006) The relationship between frontal gray matter volume and cognition varies across the healthy adult lifespan. Am J Geriatr Psychiatry 14, 823–833.

47.

Cristofori

, Zhong

, Chau

, Solomon

, Krueger

, Grafman

(2015) White and gray matter contributions to executive function recovery after traumatic brain injury. Neurology 84, 1394–1401.

48.

Sachs-Ericsson

, Blazer

(2015) The new DSM-5 diagnosis of mild neurocognitive disorder and its relation to research in mild cognitive impairment. Aging Ment Health 19, 2–12.

49.

Ganguli

, Blacker

, Blazer

, Grant

, Jeste

, Paulsen

, Petersen

, Sachdev

(2011) Classification of neurocognitive disorders in DSM-5: A work in progress. Am J Geriatr Psychiatry 19, 205–210.

50.

Kail

, Salthouse

(1994) Processing speed as a mental capacity. Acta Psychol (Amst) 86, 199–225.

51.

Corbetta

, Patel

, Shulman

(2008) The reorienting system of the human brain: From environment to theory of mind. Neuron 58, 306–324.

52.

Uncapher

, Hutchinson

, Wagner

(2011) Dissociable effects of top-down and bottom-up attention during episodic encoding. J Neurosci 31, 12613–12628.

53.

Rissman

, Wagner

(2012) Distributed representations in memory: Insights from functional brain imaging. Annu Rev Psychol 63, 101.

54.

Otten

, Henson

, Rugg

(2001) Depth of processing effects on neural correlates of memory encoding. Brain 124, 399–412.