Abstract

Cognitive slowing in individuals with multiple sclerosis (MS) has been documented by numerous studies employing explicitly timed measures in which speed of responding is an obvious focus of task performance. The present study examined information processing speed in MS patients and controls with a computerized battery of covertly timed as well as explicitly timed measures. The explicit measures were derived from two tests requiring rapid serial processing of visual stimuli, the Stroop Test and a Picture Naming Test. Covert measures were derived from the Rotated Figures Test, Remote Associates Test, and Tower of London, all tasks in which participants’ attention was drawn toward arriving at an accurate solution, and the latency with which they arrived at these solutions was timed by the computer “behind the scenes.” Significant differences in processing speed for patients and controls occurred on both types of measures, although the effect sizes were notably larger on the explicit measures.

Introduction

Early studies (e.g., Rao, Leo, Bernardin, & Unverzagt, 1991) of cognitive impairment in conjunction with multiple sclerosis (MS) indicated involvement of a variety of cognitive domains, as might be expected given the multifocal nature of this disease. However, more recent research has featured a distinct narrowing of focus on information processing speed, with several investigators concluding that slowing in the speed of information processing is the primary cognitive deficit in MS (Archibald & Fisk, 2000; DeLuca, Chelune, Tulsky, Lengenfelder, & Chiaravalloti, 2004; Demaree, DeLuca, Gaudino, & Diamond, 1999; Denney, Lynch, Parmenter, & Horne, 2004; De Sonneville et al., 2002; Kail, 1998; Kujala, Portin, Revonsuo, & Ruutiainen, 1994; Macniven et al., 2008).

The implicit assumption held by most investigators is that this deficit is a general one, extending across a wide variety of cognitive operations. The sheer consistency with which robust differences between patients and controls have been reported on a variety of measures relating to processing speed appears to offer a compelling argument for this view. These measures include simple and complex reaction time (De Sonneville et al., 2002; Kujala et al., 1994; Reicker, Tombaugh, Walker, & Freedman, 2007), the Paced Auditory Serial Addition Test (DeLuca, Johnson, & Natelson, 1993; Demaree et al., 1999; Rao, St. Aubin-Faubert, & Leo, 1989), the Symbol Digit Modalities Test (Drake et al., 2010), the Trail Making Test (DeLuca, Johnson, Beldowicz, & Natelson, 1995; Krupp, Sliwinski, Masur, Friedberg, & Coyle, 1994), the Stroop Test (Denney & Lynch, 2009; Macniven et al., 2008; Vitkovitch, Bishop, Dancey, & Richards, 2002), and the n-back procedure (Lengenfelder et al., 2006; Parmenter, Shucard, Benedict, & Shucard, 2006; Parmenter, Shucard, & Shucard, 2007). Given dissimilarities in the cognitive operations required by these tasks (e.g., working memory, numerical skill, naming), the collective findings emerging from these studies argue for a generalized deficit in MS patients’ processing speed.

On the other hand, there are obvious similarities across these measures. All of them involve explicitly timed tasks with no attempt to hide the fact that speedy performance is a crucial determinant of ones score. Also, as we have pointed out elsewhere (Bodling, Denney, & Lynch, 2008), except for reaction time measures, all of the tasks feature a rapid serial processing format. Items appear sequentially with little or no variation in the operation to be performed on each item. The operation itself is typically not very difficult, but must be executed quickly, the goal usually being to complete as many items as possible in some allotted period of time. In essence, one might argue that the evidence supporting a generalized deficit in information processing speed derives from a limited set of measures conforming to just two formats (reaction time and rapid serial processing) in which the cognitive operations are quite simple and the critical importance of speed is obvious to participants throughout the task. The goal of the present study was to extend the evaluation of information processing speed in MS patients beyond these boundaries, to determine whether the proposition of a generalized deficit in processing speed is upheld on tasks where the cognitive operations are more complex and the importance of timing is not an obvious feature of the task being performed. The tasks were designed to measure the accuracy of participants’ solutions to problems involving planning, mental rotation, and verbal association, but also offered the opportunity to “covertly” measure the length of time required for each solution. This opportunity typically arises when preparing tests for computerized administration; it is a simple matter to design the program so that response latencies to items are captured and recorded as part of the data file, but outside of the participant's awareness.

The resulting covertly timed measures bring other advantages to the evaluation of deficits in processing speed. Some investigators have pointed out that ancillary problems occurring in conjunction with MS such as visual acuity (Bruce, Bruce, & Arnett, 2007), oral-motor (Arnett, Smith, Barwick, Benedict, & Ahlstrom, 2008; Smith & Arnett, 2007), or ocular-motor defects (Bodling, Denney, & Lynch, 2008) may contribute to the apparent slowing in patients’ cognitive speed. Though such problems cannot be eliminated entirely, the usual recourse has been to try to minimize these differences through procedural modifications to explicitly timed tasks. Thus, for example, manual dexterity problems are avoided by having patients respond verbally to items of the Symbol Digit Modalities Test, although this modification is made at the expense of whatever oral-motor problems the patients may have. An advantage of covertly timed measures is that they permit an evaluation of processing speed during periods dominated almost completely by quiet contemplation when participants are presumably reflecting upon or mentally rehearsing their answer to the problem before them. These measures therefore avoid most of the confounding by any ancillary perceptual motor deficits the patients may have.

On some of the explicitly timed tests, it has been possible to systematically vary the complexity of the task and thereby examine the interaction between information processing speed and cognitive load. The divergence in processing speed between patients and controls has been shown to widen with increasing cognitive demand on both reaction time measures (De Sonneville et al., 2002; Kujala et al., 1994; Reicker et al, 2007), n-back measures (Lengenfelder et al., 2006; Parmenter et al., 2006, 2007), and the Stroop Test (Denney et al., 2004). Covertly timed measures offer similar opportunities to examine this interaction, although currently it appears only one study (Denney et al., 2004) examining planning times on Tower of London (TOL) items of varying difficulty has actually done so.

In the present study, we compared the performance of MS patients and controls on a set of explicitly and covertly timed measures. The explicit measures consisted of the Stroop Test and the Picture Naming Test (PNT), two tasks requiring rapid serial processing of information that clearly distinguish between MS patients and controls (Bodling et al., 2008). The covert measures included recordings of the length of time participants spent planning the sequence of moves they would make on separate problems of the TOL and the latencies of their correct responses to items composing the Rotated Figures Test (RFT) and the Remote Associates Test (RAT). Of these covert measures, only the TOL has been included in previous comparisons of MS patients and controls (Arnett et al., 1997; Denney et al., 2004; Foong et al., 1997). We hypothesized that the MS patients would exhibit slower information processing speed on both the explicitly timed and covertly timed measures. Also, wherever it was possible to examine systematic variation in cognitive load, we expected to find greater divergence in processing speed between patients and controls with increasing levels of cognitive demand.

Methods

Research Participants

The sample consisted of 40 patients with clinically definite MS (Poser et al., 1983) and 40 healthy controls. All the patients were under the care of the same board-certified neurologist (SGL) at the University of Kansas Medical Center, who limited recruitment to individuals with sufficient cognitive ability to provide informed consent and understand the study instructions. The following additional exclusionary criteria were used: (a) presence of any neurological disorder other than MS; (b) past or present alcohol or substance abuse; (c) visual impairment exceeding 20/50 or color blindness; (d) current use of narcotics or benzodiazepines; and (e) any exacerbation of MS symptoms in the previous 30 days. The patients (36 women, 4 men) ranged between 20 and 60 years of age (M = 44.8). Age at diagnosis ranged from 12 to 53 years (M = 34.3), and duration of diagnosis from 2 to 32 year (M = 10.7). There were 25 patients with relapsing-remitting MS and 15 with secondary progressive MS. Disability ratings based on the Expanded Disability Status Scale (Kurtzke, 1983) ranged from 0 to 7.5 (M = 3.3; Mdn = 2.5).

The healthy controls (36 women and 4 men), ranging in age from 25 to 59 (M = 40.3), were recruited from the metropolitan areas of Lawrence, Kansas and St. Louis, Missouri. They had no chronic medical problems, were following no ongoing medication regimen, and met the same exclusionary criteria as the MS patients.

Measures

Five computerized cognitive tests were administered to participants in a fixed order. The covertly timed tests (RFT, RAT, TOL) preceded the explicitly timed tests (Stroop Color-Word Test, PNT) so that any focus on speed of performance did not become apparent until after the covertly timed tests had been completed.

The Rotated Figures Test

The RFT made use of the same type of three-dimensional arrangements of blocks employed by Shepard and Metzler (1971). Each item consisted of two block arrangements presented side by side on the computer screen. The two arrangements were either the same or mirror images of one another and were rotated relative to each other between 20 and 180° around the vertical axis. During the instructional portion of the program, participants were shown four animated examples of how one block arrangement could be rotated to determine whether it did or did not matched the other. They were then administered 12 practice items, immediately followed by 54 actual items. The actual items consisted of nine items (five matching figures and four mirror-image figures) in each of six rotations (20, 40, 80, 120, 140, and 180°). Items were arranged in a randomly determined order which was the same for all participants. Participants responded by saying “yes” if the two images matched or “no” if they did not, and the experimenter immediately pressed one of two keys to record the participant's response. The computer timed the interval between the presentation of the item and the key press, recording these latencies as well as the accuracy of the responses. In addition to individual item data, the program aggregated the accuracy score for each rotation and also averaged the latency scores for the correctly answered items in which the two block arrangements matched each other (i.e., yes-correct latencies) at each rotation.

The Remote Associates Test

The RAT was adapted from the work of Mednick and Mednick (1967). Our computerized version of this test consisted of four practice items and 20 test items. Each item consisted of three words appearing side by side on the computer screen (e.g., COOKIES, SIXTEEN, HEART). The participant was instructed to think of a fourth word relating to all three of the stimulus words (e.g., SWEET). They were told they could take as long as they needed, but they should say the word out loud as soon as they thought of it. They were also permitted to say “pass,” if they were unable to think of a word and wished to skip the item. When the participant responded, the experimenter pressed a key that stopped the timing of the participant's response and replaced the stimulus words with a box where the experimenter typed in that response. The computer recorded the participant's response and the elapsed time between the presentation of the item and the occurrence of this response. The program compiled the score on the 20 items of the test (recording the total number of correct, incorrect, and passed items) and averaged the latency scores for the correctly answered items (i.e., correct latencies).

The Tower of London

The TOL is a test of planning and strategic problem solving originally described by Krikorian, Bartok, and Gay (1994). In the upper portion of the screen, the computer displayed three colored disks arranged on three pegs. In the lower portion of the screen, the computer displayed a model with the disks in a different arrangement on the pegs. The participant's task was to move the disks in the upper display so they matched the arrangement in the bottom display and to do so using a specified number of moves. Moves were made by verbal dictation, the participant stating first the color of the disk and then the number of the peg to which it was to be moved. A sample problem was used to introduce the task and acclimate the participant to the dictation method of responding; this sample problem was repeated until solved successfully. Twelve problems were then presented, graduated in difficulty from those that could be solved in two moves to those requiring a sequence of five moves. The number of moved permitted for each problem was announced prior to displaying the problem so that silence prevailed when the problem was displayed. The computer measured the length of time between the display of the problem and the participant's announcement of the color of the first disk to be moved (“planning time”). Participants were allowed three attempts to solve each problem in the specified number of moves and were awarded 3, 2, or 1 point for success on the first, second, or third attempt. The point score was summed across the 12 problems. The computer also recorded the planning time for each trial. The planning times for the initial trial of each problem (regardless of whether the problem was solved successfully during this first attempt) were averaged for the 2-move, 3-move, 4-move, and 5-move problems of the test. This measure was termed “initial planning time.”

The Stroop Color-Word Test

The Stroop Color-Word Test (Golden, 1978) consisted of three 60-s trials in which participants (a) read one of four color words (word reading), (b) named one of four colors displayed as “XXXX” (color naming), and (c) named the color of the letters used to print one of four color words (color-word naming) displayed in the center of the screen. A brief, 8-item practice set was presented before the start of each trial. Also, prior to each trial, participants were instructed as follows: “Work quickly but try not to make any mistakes. If you do make an error, try not to correct it. Just go on to the next item.” Consistent with these instructions, the examiner was trained to act like a voice-activated relay, pressing the spacebar regardless of the participant's response. The computer timed the trial and recorded the number of items completed during the trial. Errors occur rarely in this task and were not recorded in this study.

In addition to the word reading (W), color naming (C), and color-word naming (CW) scores, two combined scores were determined: (a) the sum of the word reading and color naming scores (W + C), which appears to be a particularly sensitive measure of processing speed on the Stroop (Denney & Lynch, 2009); and (b) the relative interference score, that is, (C − CW)/C, indicating the degree to which participants’ responses during the third trial were hampered by the incongruity between the words and the color of the letters. Relative interference provides a better measure of interference than the simple difference between C and CW when groups differ substantially in processing speed (Denney & Lynch, 2009).

The Picture Naming Test

The PNT was originally developed to examine the contribution of patients’ ancillary ocular-motor and oral-motor problems to measures of processing speed on tests calling for rapid serial processing of information (Bodling et al., 2008). The test uses a computerized format similar to that of the Stroop. Participants were asked to name each object or animal depicted in a series of readily recognizable, achromatic line drawings that appeared on the screen. As soon as the participant responded, the experimenter pressed the space bar to display the next picture. The test consisted of four 60-s trials, each preceded by an 8-item practice session. In the first trial (PN1), the stimuli were presented in the center of the computer screen and consisted of one of only four different pictures (bell, dog, fan, pencil), so each picture was repeated several times over the course of the trial. In the second trial (PN2), a set of 50 different pictures was used, with no picture repeated during the trial. In the third trial (PN3), the same four pictures used in trial 1 were now shown in one of nine different random locations arrayed in a 3 × 3 matrix on the screen. In the fourth trial (PN4), a different set of 50 pictures was used, with each appearing in one of the nine different random locations on the screen. Two different sets of 50 pictures were developed for the second and fourth trials, and the order of these sets was randomized and counterbalanced across participants. If they did not recognize a particular drawing, participants were permitted to say “pass,” and the next picture was displayed. The computer timed the trial and recorded the number of stimuli completed during the trial. Naming errors and passes occurred very rarely and were not analyzed.

Procedure

This study was approved by the Human Subjects Committee of the University of Kansas Medical Center, and informed consent was obtained from all participants. Patients were recruited during the course of their regular clinical appointment in the MS Clinic. Controls were initially contacted in person or by phone. If they agreed to participate, an appointment was scheduled at the participant's convenience to complete the testing either at the KU Medical Center or in the participant's home. The full testing appointment took 60–75 min.

Results

Initial Differences Between Patients and Controls

Patients and controls did not differ significantly in age. However, patients had lower ratings for education than the controls (t = 4.22, df = 78, p < .001), where education was rated on a scale from 1 to 5 (1 = completed high school education, 2 = some college, 3 = completed 4-year degree, 4 = some graduate school, and 5 = completed advanced graduate degree). Due to this difference, all analyses of the cognitive variables were first conducted with education entered as a covariate. Because education did not emerge as a significant covariate in any of these analyses, it was not included in the analyses reported below.

Explicitly Timed Measures

The findings for the explicitly timed measures of information processing speed are presented in Table 1. Except for the interference score on the Stroop, patients achieved significantly lower scores on all of these measures (all p's < .001). The relative interference score is not a measure of processing speed (Denney & Lynch, 2009). Instead, it reflects the amount of distraction resulting from incongruity between words and colors and is effectively corrected for differences in processing speed. The effect sizes (Cohen's d) for the differences in processing speed assessed with explicit measures were large, ranging from 1.28 (CW) to 2.16 (W + C).

Table 1.

Comparisons between MS patients and controls on explicitly timed measures

Cognitive measure MS patients (M [SD]) Controls (M [SD]) t-value (df = 78) p-value Effect size (Cohen's d
Stroop Test 
 Word reading (W) 73.05 (7.78) 88.28 (7.76) 8.77 <.001 1.95 
 Color naming (C) 60.68 (7.95) 74.35 (5.46) 8.97 <.001 2.00 
 Color-word naming (CW) 43.20 (8.41) 53.05 (6.95) 5.71 <.001 1.28 
 Sum: W + C 133.73 (14.75) 162.63 (11.82) 9.67 <.001 2.16 
 Relative interference 28.83 (9.36) 28.63 (7.77) 0.10 .920 — 
Picture Naming Test 
 Trial 1 (PN1) 60.22 (7.01) 72.03 (5.64) 8.30 <.001 1.86 
 Trial 2 (PN2) 42.95 (8.93) 52.95 (5.13) 6.14 <.001 1.51 
 Trial 3 (PN3) 56.10 (6.61) 67.75 (4.76) 9.05 <.001 2.02 
 Trial 4 (PN4) 41.18 (8.31) 52.88 (4.58) 7.80 <.001 1.74 
 Average: Trials 1–4 50.11 (7.09) 61.40 (4.37) 8.57 <.001 1.92 
Cognitive measure MS patients (M [SD]) Controls (M [SD]) t-value (df = 78) p-value Effect size (Cohen's d
Stroop Test 
 Word reading (W) 73.05 (7.78) 88.28 (7.76) 8.77 <.001 1.95 
 Color naming (C) 60.68 (7.95) 74.35 (5.46) 8.97 <.001 2.00 
 Color-word naming (CW) 43.20 (8.41) 53.05 (6.95) 5.71 <.001 1.28 
 Sum: W + C 133.73 (14.75) 162.63 (11.82) 9.67 <.001 2.16 
 Relative interference 28.83 (9.36) 28.63 (7.77) 0.10 .920 — 
Picture Naming Test 
 Trial 1 (PN1) 60.22 (7.01) 72.03 (5.64) 8.30 <.001 1.86 
 Trial 2 (PN2) 42.95 (8.93) 52.95 (5.13) 6.14 <.001 1.51 
 Trial 3 (PN3) 56.10 (6.61) 67.75 (4.76) 9.05 <.001 2.02 
 Trial 4 (PN4) 41.18 (8.31) 52.88 (4.58) 7.80 <.001 1.74 
 Average: Trials 1–4 50.11 (7.09) 61.40 (4.37) 8.57 <.001 1.92 

The three trials of the Stroop can be thought of as varying in cognitive load and thus a two-way analysis can be used to examine the interaction between group and trial. For this analysis, the number of items completed on each 60 s trial was converted to the number of seconds per item so that the resulting graph would be comparable to those generated for covertly timed measures presented below. These reciprocal scores were analyzed with a 2 (Group) × 3 (Trial) mixed factorial analysis of variance. Significant main effects were found for trial (F = 287.59, df = 2,77, p < .001) and group (F = 62.72, df = 1,78, p < .001), along with a significant group × trial interaction (F = 5.99, df = 2,77, p = .004). Response times lengthened across the three successive trials of the Stroop, but the impact of this increasing cognitive demand was greater on patients’ response times than on those for controls (Fig. 1).

Fig. 1.

Stroop Test: Response times (s) per completed item by trial.

Fig. 1.

Stroop Test: Response times (s) per completed item by trial.

The stimuli used in the four trials of the PNT varied along two bipolar dimensions: They were either repeated multiple times (Trials 1 and 3) or unique (Trials 2 and 4) throughout the trial; and they were presented either in the center of the screen (Trials 1 and 2) or distributed randomly in one of nine locations on the screen (Trials 3 and 4). A 2 (Group) × 2 (Repetition) × 2 (Distribution) mixed factorial analysis of variance performed on the scores for the four trials revealed significant main effects for group (F = 73.42, df = 1,79, p < .001), repetition (F = 101.63, df = 1,78, p < .001), and distribution (F = 74.75, df = 1,78, p < .001), and a significant interaction between repetition and distribution (F = 50.88, df = 1,78, p < .001), but no significant interactions involving group. Presentation of unique pictures in random locations on the screen throughout the trial clearly resulted in lower scores, but in contrast to the variation in cognitive load examined in the preceding analysis of the Stroop trials, these variations in the stimulus parameters of the PNT impacted the performance of patients and controls equally.

Covertly Timed Measures

Comparisons between patients and controls in terms of both their accuracy scores and their response latencies on the covertly timed measures of processing speed are presented in Table 2. Since the response to each item of the RFT was dichotomous, a signal detection analysis was used to compute accuracy (d') on this test. The only difference in accuracy that attained significance occurred on the RFT; here, patients made fewer correct responses than controls (t = 3.37, df = 78, p=.001). On the other hand, significantly slower processing speeds in MS patients were evident on all three covert measures, with the effect sizes for these differences ranging from 0.65 on the RAT to 0.74 on the RFT.

Table 2.

Comparisons between MS patients and controls on covertly timed measures

Cognitive measure MS Patients (M [SD]) Controls (M [SD]) t-value (df = 78) p-value Effect size (Cohen's d
Rotated Figures Test 
 Accuracy (d') 1.74 (1.33) 2.79 (1.45) 3.37 .001 0.75 
 Yes-correct latencies 10.37 (4.47) 7.39 (2.88) 3.54 .001 0.79 
Remote Associates Test 
 Total correct 9.38 (3.47) 10.45 (2.86) 1.51 .134 — 
 Correct latencies 13.02 (4.81) 10.23 (3.63) 2.92 .005 0.65 
Tower of London 
 Point score 31.65 (3.85) 33.00 (1.97) 1.97 .052 — 
 Initial planning times 18.86 (7.51) 14.22 (5.35) 2.92 .005 0.71 
Cognitive measure MS Patients (M [SD]) Controls (M [SD]) t-value (df = 78) p-value Effect size (Cohen's d
Rotated Figures Test 
 Accuracy (d') 1.74 (1.33) 2.79 (1.45) 3.37 .001 0.75 
 Yes-correct latencies 10.37 (4.47) 7.39 (2.88) 3.54 .001 0.79 
Remote Associates Test 
 Total correct 9.38 (3.47) 10.45 (2.86) 1.51 .134 — 
 Correct latencies 13.02 (4.81) 10.23 (3.63) 2.92 .005 0.65 
Tower of London 
 Point score 31.65 (3.85) 33.00 (1.97) 1.97 .052 — 
 Initial planning times 18.86 (7.51) 14.22 (5.35) 2.92 .005 0.71 

Both the RFT and the TOL featured systematic variations in cognitive demand so that a two-way analysis could be used to examine the interaction between group and problem difficulty. A 2 (Group) × 6 (Rotation) mixed factorial analysis of variance of the yes-correct latencies on the RFT yielded significant main effects for rotation (F = 27.40; df = 5,74; p < .001) and group (F = 11.04; df = 1,78; p = .001). Response latencies lengthened as the degree of rotation between the two figures increased, and, regardless of rotation, patients responded more slowly than controls. However, the group × rotation interaction was not significant (F = 1.68; df = 5,74; p = .15); the increase in demand occasioned by greater rotation impacted patients’ latencies the same as controls (Fig. 2).

Fig. 2.

RFT: Latencies for yes-correct responses by degree of rotation.

Fig. 2.

RFT: Latencies for yes-correct responses by degree of rotation.

Initial planning times on the TOL were analyzed in a similar fashion, using a 2 (Group) × 4 (Problem Type) mixed factorial analysis of covariance. Significant main effects occurred for problem type (F = 57.25; df = 3,76; p < .001) and group (F = 9.41; df = 1,78; p = .003) because planning times lengthened as the test progressed from 2-move to 5-move problems and because patients exhibited longer planning times than controls across all problems. Here again, however, the group × problem type interaction did not attain significance (F = 1.92; df = 3,76; p = .13). This interaction is illustrated in Fig. 3.

Fig. 3.

TOL: Initial planning times by problem type.

Fig. 3.

TOL: Initial planning times by problem type.

Correlations between Explicitly Timed and Covertly Timed Measures of Processing Speed

Bivariate correlations were computed separately for each group and then combined using Fisher's method of transforming the correlation coefficients into z-values. The resulting correlations between the various explicit measures of processing speed ranged from .57 to .94 (all p's < .001) and those between the covert measures of processing speed ranged from .35 to .50 (all p's ≤ .001). Relative to these correlations within categories, correlations between explicit and covert measures were smaller (Table 3). A principal components analysis performed on the full array of processing speed measures resulted in two factors. The first factor (eigenvalue = 6.0) appeared to represent general information processing speed. The seven overtly timed measures loaded substantially on this factor (0.81–0.96); the loadings for the three covertly timed measures were notably lower—ranging from −0.32 (TOL) to −0.43 (RFT). The covertly timed measures had higher loadings (0.61–0.77) on the second factor (eigenvalue = 1.5); the loadings of the explicitly timed measures on the second factor were negligible (0.07–0.17).

Table 3.

Correlations between explicitly and covertly timed measures of processing speed

Cognitive measure Rotated Figures Test: Yes-correct latency
 
Remote Associates Test: Correct latency
 
Tower of London: Initial planning time
 
R-value p-value R-value p-value R-value p-value 
Stroop Test 
 Word reading (W) −.215 .055 −.211 .060 −.157 .165 
 Color naming (C) −.260 .020 −.208 .064 −.195 .083 
 Color-word naming (CW) −.318 .004 −.259 .020 −.151 .182 
 Sum: W + C −.246 .028 −.219 .051 −.182 .105 
Picture Naming Test 
 Trial 1 (PN1) −.315 .004 −.206 .067 −.194 .084 
 Trial 2 (PN2) −.202 .073 −.370 .007 −.172 .127 
 Trial 3 (PN3) −.312 .005 −.265 .018 −.167 .138 
 Trial 4 (PN4) −.240 .032 −.388 <.001 −.194 .084 
 Average: Trials 1–4 −.282 .011 −.327 .003 −.193 .086 
Cognitive measure Rotated Figures Test: Yes-correct latency
 
Remote Associates Test: Correct latency
 
Tower of London: Initial planning time
 
R-value p-value R-value p-value R-value p-value 
Stroop Test 
 Word reading (W) −.215 .055 −.211 .060 −.157 .165 
 Color naming (C) −.260 .020 −.208 .064 −.195 .083 
 Color-word naming (CW) −.318 .004 −.259 .020 −.151 .182 
 Sum: W + C −.246 .028 −.219 .051 −.182 .105 
Picture Naming Test 
 Trial 1 (PN1) −.315 .004 −.206 .067 −.194 .084 
 Trial 2 (PN2) −.202 .073 −.370 .007 −.172 .127 
 Trial 3 (PN3) −.312 .005 −.265 .018 −.167 .138 
 Trial 4 (PN4) −.240 .032 −.388 <.001 −.194 .084 
 Average: Trials 1–4 −.282 .011 −.327 .003 −.193 .086 

Discussion

The present investigation demonstrates that MS patients perform more slowly relative to controls on a wider variety of cognitive measures than those employed by other studies, thereby adding to the evidence of a general deficit in processing speed that is quite independent of both the cognitive operations involved in the tasks and the participant's awareness of the speeded nature of the test. However, this conclusion cannot be made unequivocally because of three inconsistencies in the present findings: (a) effect sizes were larger for the explicit measures (1.28–2.16) than for the covert measures (0.65–0.74); (b) the complexity effect was evident on an explicit measure (i.e., the Stroop), but not on the covert measures; and (c) the correlations between explicit and covert measures were modest and accordingly, the principal component analysis assigned these measures to separate factors instead of yielding a single factor. The following discussion is framed around these three inconsistencies.

Differences in Effect Size

It might be tempting to attribute the larger effect sizes on explicit tests to the added contribution of ancillary deficits in patients’ sensory and motor functioning. Some investigators (Arnett et al., 2008; Bruce et al., 2007; Smith & Arnett, 2007) have argued deficits of this kind substantially inflate the apparent differences in processing speed between MS patients and controls. Modifications to the format of explicitly timed tests can never be entirely successful at eliminating these potential confounds. However, latencies on the covertly timed tests are recorded during contemplative intervals between the presentation of each problem and the initiation of its solution, at a time when sensory and motor functioning is minimal. Might one consider the effect sizes for these covert measures as the more accurate index of the deficits in patients’ central processing speed and the disparity in effect sizes between explicit and covert measures as an index of the inflation resulting from ancillary problems?

While tempting, this perspective is challenged by the results of the PNT that indicate ancillary problems make a considerably smaller contribution to patients’ processing speed deficits on explicitly timed measures than other investigators have suggested. Variations in the trials of the PNT designed to amplify oral-motor and ocular-motor demands of the task succeeded in slowing performance for all participants, but the impact on patients was no different than that of controls. This result completely replicates the findings of an earlier study (Bodling et al., 2008). Thus, we believe that the disparity in effect size obtained in the present study cannot simply be attributed to the contribution made by ancillary problems.

A more likely explanation involves variability. Effect size statistics are sensitive to variability as well as to mean differences in performance between groups; the pooled standard deviation across all participants is used in the denominator of Cohen's effect size statistic. Explicit instruction to perform as quickly as possible likely results in substantially less variability in response speed between participants than when this aspect of performance is left to participants’ own determination as it is on covertly timed tests. The greater variability on covert measures could account for the smaller effect sizes observed on these measures.

Disparity in Complexity Effect

Other studies (Denney et al., 2004; Lengenfelder et al., 2006; Parmenter et al., 2006, 2007; Reicker et al, 2007) have shown that the differences in processing speed between patients and controls widen with increasing cognitive demand. Analysis of performance across the trials of the explicitly timed Stroop Test succeeded in replicating this complexity effect (Reicker et al., 2007) in the present study. However, both the RFT and the TOL also permitted evaluations of processing speed under conditions of varying cognitive load, and in neither instance did the interaction attain significance. Here again, the greater variability on covert measures of processing speed may have contributed to this failure—along with a couple of other limitations. There was a general restriction in the range of cognitive demand encompassed by each of these covert tests and likewise a limitation on the number of items assessing processing speed at each level of cognitive demand. Figs 1 and 2 both reveal a slight divergence in the plots for patients and controls as greater cognitive demand was imposed. Due to the limitations, however, the study may have simply lacked sufficient power to detect the so-called complexity effect.

One Processing Speed or Two

The most serious challenge to the conclusion of a general deficit in processing speed evident on both overt and covert measures involves the modest correlation between these measures. A useful way of conceptualizing performance on these two types of measures might be to consider the difference between sprinting and jogging. Although both are forms of running, the nature of these two activities is quite different. Sprinting is more purely a matter of speed and therefore more readily seen as an “ability,” whereas jogging is more comfortably viewed as a kind of personal “tempo,” one affording a greater range of choice on the part of the runner. The motivations that invest sprinting are different and more restricted than those of jogging; a greater variety of factors may affect the latter. Nevertheless, the speed with which one sprints will bear some relationship with the speed with which one jogs, and furthermore when aging comes to diminish one, it also diminishes the other. Explicitly timed measures evaluate “cognitive speed,” whereas covertly timed measures assess something more akin to “cognitive tempo.” MS has an evident and dramatic impact on both, and this is the sense in which we maintain that the full array of measures in the present study reflect a general deficit in processing speed.

Within this context, the findings from the principal components analysis are less disturbing. The eigenvalue for the first factor (6.0) was substantially higher than the second (1.5), and although the three covert measures had higher loadings on the second factor, they also loaded on the first factor. The converse was not true: The seven overt measures loaded heavily on the first factor, but only negligibly on the second (0.07–0.17). It would seem the first factor provided a general index of information processing speed and the second encompassed a broader array of motivational elements more squarely involved in cognitive tempo. One example might be the participants’ trade-off between accuracy and speed. Patients’ awareness of their disease may lead them to harbor greater concerns than controls over their performance on neuropsychological tests. Such concerns might prompt patients to perform more deliberately on covert measures of processing speed, taking more time to labor over their answers. These additional elements may contribute to greater variability on covert measures and therefore account for lower effect sizes and decreased likelihood of having sufficient power to detect complexity hypothesis lying in the data.

Chiaravalloti, Chistodoulou, Demaree, and DeLuca (2003) have drawn a distinction between simple and complex measures of processing speed and have shown that MS patients perform more slowly than controls on both types of measures. Using Chiaravalloti's dichotomy, all of the covertly timed measures used in the present study would have to be considered the measures of complex processing speed. However, a wider distinction is in play here. Beyond mere complexity, the covertly timed measures assess differences in processing speed when participants are attending to the accuracy of their eventual response rather than the speed with which it is achieved. They also assess these differences in processing speed when confounding due to ancillary motor deficits is minimized. Even under the diverse conditions of the study, it is evident that MS patients exhibit significant deficits in their speed of information processing.

Conflict of Interest

None declared.

Acknowledgements

This research was unfunded and the authors have no financial or other conflicts of interests to report. A paper based on this research was presented on September 18, 2008 at the World Congress for Treatment and Research in Multiple Sclerosis, Montreal, Canada.

References

Archibald
C. J.
Fisk
J. D.
Information processing efficiency in patients with multiple sclerosis
Journal of Clinical and Experimental Neuropsychology
 , 
2000
, vol. 
22
 (pg. 
686
-
701
)
Arnett
P. A.
Rao
S. M.
Grafman
J.
Bernardin
L.
Lucetta
T.
Binder
J. R.
, et al.  . 
Executive functions in multiple sclerosis: An analysis of temporal ordering, semantic encoding, and planning abilities
Neuropsychology
 , 
1997
, vol. 
11
 (pg. 
535
-
544
)
Arnett
P. A.
Smith
M. M.
Barwick
F. H.
Benedict
R. H. B.
Ahlstrom
B. P.
Oralmotor slowing in multiple sclerosis: Relationship to neuropsychological tasks requiring an oral response
Journal of the International Neuropsychological Society
 , 
2008
, vol. 
14
 (pg. 
454
-
462
)
Bodling
A. M.
Denney
D. R.
Lynch
S. G.
Rapid serial processing in patients with multiple sclerosis: The role of peripheral deficits
Journal of the International Neuropsychological Society
 , 
2008
, vol. 
14
 (pg. 
646
-
650
)
Bruce
J. M.
Bruce
A. S.
Arnett
P. A.
Mild visual acuity disturbances are associated with performance on tests of complex attention in MS
Journal of the International Neuropsychological Society
 , 
2007
, vol. 
13
 (pg. 
544
-
548
)
Chiaravalloti
N. D.
Chistodoulou
C.
Demaree
H. A.
DeLuca
J.
Differentiating simple versus complex processing speed: Influence on new learning and memory performance
Journal of Clinical and Experimental Neuropsychology
 , 
2003
, vol. 
25
 (pg. 
489
-
501
)
DeLuca
J.
Chelune
G. J.
Tulsky
D. S.
Lengenfelder
J.
Chiaravalloti
N. D.
Is speed of processing or working memory the primary information processing deficit in multiple sclerosis?
Journal of Clinical and Experimental Neuropsychology
 , 
2004
, vol. 
26
 (pg. 
550
-
562
)
DeLuca
J.
Johnson
S. K.
Beldowicz
D.
Natelson
B. H.
Neuropsychological impairments in chronic fatigue syndrome, multiple sclerosis, and depression
Journal of Neurology, Neurosurgery, and Psychiatry
 , 
1995
, vol. 
58
 (pg. 
38
-
43
)
DeLuca
J.
Johnson
S. K.
Natelson
B. H.
Information processing efficiency in chronic fatigue syndrome and multiple sclerosis
Archives of Neurology
 , 
1993
, vol. 
50
 (pg. 
301
-
304
)
Demaree
H. A.
DeLuca
J.
Gaudino
E. A.
Diamond
B. J.
Speed of information processing as a key deficit in multiple sclerosis: Implications for rehabilitation
Journal of Neurology, Neurosurgery, and Psychiatry
 , 
1999
, vol. 
67
 (pg. 
661
-
663
)
Denney
D. R.
Lynch
S. G.
The impact of multiple sclerosis on patients’ performance on the Stroop Test: Processing speed vs. interference
Journal of the International Neuropsychological Society
 , 
2009
, vol. 
15
 (pg. 
451
-
458
)
Denney
D. R.
Lynch
S. G.
Parmenter
B. A.
Horne
N.
Cognitive impairment in relapsing and primary progressive multiple sclerosis: Mostly a matter of speed
Journal of the International Neuropsychological Society
 , 
2004
, vol. 
10
 (pg. 
948
-
956
)
De Sonneville
L. M. J.
Boringa
J. B.
Reuling
I. E. W.
Lazeron
R. H. C.
Adèr
H. J.
Polman
C. H.
Information processing characteristics in subtypes of multiple sclerosis
Neuropsychologia
 , 
2002
, vol. 
40
 (pg. 
1751
-
1765
)
Drake
A. S.
Weinstock-Guttman
B.
Morrow
S. A.
Hojnacki
D.
Munschauer
F. E.
Benedict
R. H. B.
Psychometrics and normative data for the Multiple Sclerosis Functional Composite: Replacing the PASAT with the Symbol Digit Modalities Test
Multiple Sclerosis
 , 
2010
, vol. 
16
 (pg. 
228
-
237
)
Foong
J.
Rozewicz
L.
Quaghebeur
G.
Davie
C. A.
Kartsounis
L. D.
Thompson
A. J.
, et al.  . 
Executive function in multiple sclerosis: The role of frontal lobe pathology
Brain
 , 
1997
, vol. 
120
 (pg. 
15
-
26
)
Golden
C. J.
The Stroop Color and Word Test
 , 
1978
Wood Dale, IL
Stoelting Company
Kail
R.
Speed of information processing in patients with multiple sclerosis
Journal of Clinical and Experimental Neuropsychology
 , 
1998
, vol. 
20
 (pg. 
98
-
106
)
Krikorian
R.
Bartok
J.
Gay
N.
Tower of London procedure: A standard method and developmental data
Journal of Clinical and Experimental Neuropsychology
 , 
1994
, vol. 
16
 (pg. 
840
-
850
)
Krupp
L. B.
Sliwinski
M.
Masur
D. M.
Friedberg
F.
Coyle
P. K.
Cognitive functioning and depression in patients with chronic fatigue syndrome and multiple sclerosis
Archives of Neurology
 , 
1994
, vol. 
51
 (pg. 
705
-
710
)
Kujala
P.
Portin
R.
Revonsuo
A.
Ruutiainen
J.
Automatic and controlled information processing in multiple sclerosis
Brain
 , 
1994
, vol. 
117
 (pg. 
1115
-
1126
)
Kurtzke
J. F.
Rating neurologic impairment in multiple sclerosis: An expanded disability status scale (EDSS)
Neurology
 , 
1983
, vol. 
33
 (pg. 
1444
-
1452
)
Lengenfelder
J.
Bryant
D.
Diamond
B. J.
Kalmar
J. H.
Moore
N. B.
DeLuca
J.
Processing speed interacts with working memory efficiency in multiple sclerosis
Archives of Clinical Neuropsychology
 , 
2006
, vol. 
21
 (pg. 
229
-
238
)
Macniven
J. A.
Davis
C.
Ho
M. Y.
Bradshaw
C. M.
Szabadi
E.
Constantinescu
C. S.
Stroop performance in multiple sclerosis: Information processing, selective attention, or executive functioning?
Journal of the International Neuropsychological Society
 , 
2008
, vol. 
14
 (pg. 
805
-
814
)
Mednick
S. A.
Mednick
M. T.
Examiner's Manual for the Remote Associates Test
 , 
1967
Boston
Houghton Mifflin
Parmenter
B. A.
Shucard
J. L.
Benedict
R. H. B.
Shucard
D. W.
Working memory deficits in multiple sclerosis: Comparison between the n-back task and the Paced Auditory Serial Addition Test
Journal of the International Neuropsychological Society
 , 
2006
, vol. 
12
 (pg. 
677
-
687
)
Parmenter
B. A.
Shucard
J. L.
Shucard
D. W.
Information processing deficits in multiple sclerosis: A matter of complexity
Journal of the International Neuropsychological Society
 , 
2007
, vol. 
13
 (pg. 
417
-
423
)
Poser
C. M.
Paty
D. W.
Scheinberg
L.
McDonald
W. I.
Davis
F. A.
Ebers
G. C.
, et al.  . 
New diagnostic criteria for multiple sclerosis: Guidelines for research protocols
Annals of Neurology
 , 
1983
, vol. 
13
 (pg. 
227
-
231
)
Rao
S. M.
Leo
G. J.
Bernardin
L.
Unverzagt
F.
Cognitive dysfunction in multiple sclerosis. I. Frequency, patterns, and prediction
Neurology
 , 
1991
, vol. 
41
 (pg. 
685
-
691
)
Rao
S. M.
St. Aubin-Faubert
P.
Leo
G. J.
Information processing speed in patients with multiple sclerosis
Journal of Experimental and Clinical Neuropsychology
 , 
1989
, vol. 
11
 (pg. 
471
-
477
)
Reicker
L. I.
Tombaugh
T. N.
Walker
L.
Freedman
M. S.
Reaction time: An alternative method for assessing the effects of multiple sclerosis on information processing speed
Archives of Clinical Neuropsychology
 , 
2007
, vol. 
22
 (pg. 
655
-
664
)
Shepard
R. N.
Metzler
J.
Mental rotation of three-dimensional objects
Science
 , 
1971
, vol. 
171
 (pg. 
701
-
703
)
Smith
M. M.
Arnett
P. A.
Dysarthria predicts poorer performance on cognitive tasks required speeded oral response in an MS population
Journal of Clinical and Experimental Neuropsychology
 , 
2007
, vol. 
29
 (pg. 
804
-
812
)
Vitkovitch
M.
Bishop
S.
Dancey
C.
Richards
A.
Stroop interference and negative priming in patients with multiple sclerosis
Neuropsychologica
 , 
2002
, vol. 
40
 (pg. 
1560
-
1574
)