Response monitoring in math-anxious individuals in an arithmetic task.

We examine whether math anxiety is related to altered response monitoring in an arithmetic task. Response-locked event-related brain potentials (ERPs) were evaluated in 23 highly (HMA) and 23 low math-anxious (LMA) individuals while they performed an arithmetic verification task. We focused on two widely studied ERPs elicited during error processing: error-related negativity (ERN) and error positivity (Pe). Correct-related negativity (CRN), an ERP elicited after a correct response, was also studied. The expected ERN following errors was found, but groups did not differ in its amplitude. Importantly, LMA individuals showed less negative CRN and more positive Pe amplitudes than their more anxious peers, suggesting more certainty regarding response accuracy and better adaptive behavioral adjustment after committing errors in an arithmetic task in the LMA group. The worse control over response performance and less awareness of correct responses in the HMA group might reduce their ability to 'learn from errors'.

Proactive control of attention in math-anxious individuals.

Attentional control deficit has been proposed as one of the reasons for lower arithmetical performance in people with high math anxiety (HMA). Previous research trying to discern whether this deficit concerned proactive or reactive use of attentional control has been criticised because the methodologies used were mostly suited to investigating reactive control only. The aim of this study was to investigate proactive control in HMA individuals in a classical Stroop task. Twenty HMA and 20 low math-anxious individuals (LMA) named the ink colour in which congruent and incongruent colour words as well as X strings (neutral condition) were presented. The HMA group was slower than their LMA peers in the congruent and incongruent conditions only. Furthermore, HMA individuals showed a higher interference effect. Last, only LMA participants showed a facilitatory effect of the congruent condition. These results are interpreted as indicating the presence in the HMA individuals of a task conflict between the task to perform (ink naming) and an irrelevant task triggered by the stimuli (word reading). Task conflict is evident only when proactive control, responsible for maintaining the current goals, is too weak to solve the competition between tasks. Therefore, this study confirms that HMA individuals find it difficult to implement attention proactively.

Processing of Ordinal Information in Math-Anxious Individuals.

This study aimed to investigate whether the ordinal judgments of high math-anxious (HMA) and low math-anxious (LMA) individuals differ. Two groups of 20 participants with extreme scores on the Shortened Mathematics Anxiety Rating Scale (sMARS) had to decide whether a triplet of numbers was presented in ascending order. Triplets could contain one-digit or two-digit numbers and be formed by consecutive numbers (counting condition), numbers with a constant distance of two or three (balanced) or numbers with variable distances between them (neutral). All these triplets were also presented unordered: sequence order in these trials could be broken at the second (D2) or third (D3) number. A reverse distance effect (worse performance for ordered balanced than for counting trials) of equal size was found in both anxiety groups. However, HMA participants made more judgment errors than their LMA peers when they judged one-digit counting ordered triplets. This effect was related to worse performance of HMA individuals on a symmetry span test and might be related to group differences on working memory. Importantly, HMAs were less accurate than LMA participants at rejecting unordered D2 sequences. This result is interpreted in terms of worse cognitive flexibility in HMA individuals.

Number line estimation in highly math-anxious individuals.

In this study, we aimed to investigate the difficulties highly math-anxious individuals (HMA) may face when having to estimate a number's position in a number line task. Twenty-four HMA and 24 low math-anxiety (LMA) individuals were presented with four lines with endpoints 0-100, 0-1,000, 0-100,000, and 267-367 on a computer monitor on which they had to mark the correct position of target numbers using the mouse. Although no differences were found between groups in the frequency of their best-fit model, which was linear for all lines, the analysis of slopes and intercepts for the linear model showed that the two groups differed in performance on the less familiar lines (267-367 and 0-100,000). Lower values for the slope and higher values for the intercept were found in the HMA group, suggesting that they tended to overestimate small numbers and underestimate large numbers on these non-familiar lines. Percentage absolute error analyses confirmed that HMA individuals were less accurate than their LMA counterparts on these lines, although no group differences were found in response time. These results indicate that math anxiety is related to worse performance only in the less familiar and more difficult number line tasks. Therefore, our data challenge the idea that HMA individuals might have less precise numerical representations and support the anxiety-complexity effect posited by Ashcraft and colleagues.

Math anxiety: A review of its cognitive consequences, psychophysiological correlates, and brain bases.

A decade has passed since the last published review of math anxiety, which was carried out by Ashcraft and Ridley (2005). Given the considerable interest aroused by this topic in recent years and the growing number of publications related to it, the present article aims to provide a full and updated review of the field, ranging from the initial studies of the impact of math anxiety on numerical cognition, to the latest research exploring its electrophysiological correlates and brain bases from a cognitive neuroscience perspective. Finally, this review describes the factors and mechanisms that have been claimed to play a role in the origins and/or maintenance of math anxiety, and it examines in detail the main explanations proposed to account for the negative effects of math anxiety on performance: competition for working memory resources, a deficit in a low-level numerical representation, and inhibition/attentional control deficit.

Attentional bias in high math-anxious individuals: evidence from an emotional Stroop task.

Attentional bias toward threatening or emotional information is considered a cognitive marker of anxiety, and it has been described in various clinical and subclinical populations. This study used an emotional Stroop task to investigate whether math anxiety is characterized by an attentional bias toward math-related words. Two previous studies failed to observe such an effect in math-anxious individuals, although the authors acknowledged certain methodological limitations that the present study seeks to avoid. Twenty high math-anxious (HMA) and 20 low math-anxious (LMA) individuals were presented with an emotional Stroop task including math-related and neutral words. Participants in the two groups did not differ in trait anxiety or depression. We found that the HMA group showed slower response times to math-related words than to neutral words, as well as a greater attentional bias (math-related - neutral difference score) than the LMA one, which constitutes the first demonstration of an attentional bias toward math-related words in HMA individuals.

Processing of multi-digit additions in high math-anxious individuals: psychophysiological evidence.

We investigated the time course of neural processing of multi-digit additions in high- (HMA) and low-math anxious (LMA) individuals. Seventeen HMA and 17 LMA individuals were presented with two-digit additions and were asked to perform a verification task. Behavioral data showed that HMA individuals were slower and more error prone than their LMA peers, and that incorrect solutions were solved more slowly and less accurately than correct ones. Moreover, HMA individuals tended to need more time and commit more errors when having to verify incorrect solutions than correct ones. ERPs time-locked to the presentation of the addends (calculation phase) and to the presentation of the proposed solution (verification phase) were also analyzed. In both phases, a P2 component of larger amplitude was found for HMA individuals than for their LMA peers. Because the P2 component is considered to be a biomarker of the mobilization of attentional resources toward emotionally negative stimuli, these results suggest that HMA individuals may have invested more attentional resources both when processing the addends (calculation phase) and when they had to report whether the proposed solution was correct or not (verification phase), as compared to their LMA peers. Moreover, in the verification phase, LMA individuals showed a larger late positive component (LPC) for incorrect solutions at parietal electrodes than their HMA counterparts. The smaller LPC shown by HMA individuals when verifying incorrect solutions suggests that these solutions may have been appeared more plausible to them than to their LMA counterparts.

Reactive recruitment of attentional control in math anxiety: an ERP study of numeric conflict monitoring and adaptation.

This study uses event-related brain potentials (ERPs) to investigate the electrophysiological correlates of numeric conflict monitoring in math-anxious individuals, by analyzing whether math anxiety is related to abnormal processing in early conflict detection (as shown by the N450 component) and/or in a later, response-related stage of processing (as shown by the conflict sustained potential; Conflict-SP). Conflict adaptation effects were also studied by analyzing the effect of the previous trial's congruence in current interference. To this end, 17 low math-anxious (LMA) and 17 high math-anxious (HMA) individuals were presented with a numerical Stroop task. Groups were extreme in math anxiety but did not differ in trait or state anxiety or in simple math ability. The interference effect of the current trial (incongruent-congruent) and the interference effect preceded by congruence and by incongruity were analyzed both for behavioral measures and for ERPs. A greater interference effect was found for response times in the HMA group than in the LMA one. Regarding ERPs, the LMA group showed a greater N450 component for the interference effect preceded by congruence than when preceded by incongruity, while the HMA group showed greater Conflict-SP amplitude for the interference effect preceded by congruence than when preceded by incongruity. Our study showed that the electrophysiological correlates of numeric interference in HMA individuals comprise the absence of a conflict adaptation effect in the first stage of conflict processing (N450) and an abnormal subsequent up-regulation of cognitive control in order to overcome the conflict (Conflict-SP). More concretely, our study shows that math anxiety is related to a reactive and compensatory recruitment of control resources that is implemented only when previously exposed to a stimuli presenting conflicting information.

Abnormal error monitoring in math-anxious individuals: evidence from error-related brain potentials.

This study used event-related brain potentials to investigate whether math anxiety is related to abnormal error monitoring processing. Seventeen high math-anxious (HMA) and seventeen low math-anxious (LMA) individuals were presented with a numerical and a classical Stroop task. Groups did not differ in terms of trait or state anxiety. We found enhanced error-related negativity (ERN) in the HMA group when subjects committed an error on the numerical Stroop task, but not on the classical Stroop task. Groups did not differ in terms of the correct-related negativity component (CRN), the error positivity component (Pe), classical behavioral measures or post-error measures. The amplitude of the ERN was negatively related to participants' math anxiety scores, showing a more negative amplitude as the score increased. Moreover, using standardized low resolution electromagnetic tomography (sLORETA) we found greater activation of the insula in errors on a numerical task as compared to errors in a non-numerical task only for the HMA group. The results were interpreted according to the motivational significance theory of the ERN.

Processing false solutions in additions: differences between high- and lower-skilled arithmetic problem-solvers.

This paper focuses on the capacity to solve numerical incongruities in high- and lower-skilled arithmetic problem-solvers by investigating event-related brain potentials elicited by incorrect solutions to additions. Fifteen high-skill and fifteen low-skill individuals were presented with simple addition problems in a verification task. The proposed solution was manipulated by presenting correct solutions and incorrect solutions very close to the correct ones. Incorrect solutions elicited a negative component followed by a late positive component (LPC/P3b), whose amplitude was smaller for the low-skill group than for the high-skill group. Because the LPC/P3b amplitude has been taken as an indicator of the plausibility of the stimulus, this result suggests that incorrect solutions close to the correct ones appear more plausible to low-skilled individuals than to their high-skilled counterparts. This result is interpreted in terms of differences in the strength of association between problems and potential solutions depending on arithmetical skill.

Effects of training on the arithmetic problem-size effect: an event-related potential study.

This study uses event-related brain potentials to examine the extent to which training in arithmetic calculation increases problem-solution associative strength and reduces the problem-size effect. Participants were presented with a classic equality verification task in two sessions. The problem size was manipulated by using small, medium and large problems. Since the problem-size effect has been related to a modulation of a late positive slow wave and explained in terms of differential frequency of use with problems, practice was expected to modify the amplitude of this positive component. Results showed that the amplitude of the positive slow wave increased with problem size and decreased with practice, suggesting that practice is an important determinant of the problem-size effect.

Problem size effect and processing strategies in mental arithmetic.

We investigated the event-related brain potentials elicited by arithmetical operations whose solution requires direct memory retrieval or non-retrieval strategies. The problem size effect--the increment in reaction time for arithmetical problems with large operands--appears to be due to the selective use of non-retrieval procedures, and studies with event-related potentials have demonstrated an amplitude modulation of a late positive slow wave (range between 400-800 ms) related to the problem size effect. Two arithmetic operations (additions and subtractions) and three levels of problem size (adding or subtracting 2, 4 or 6) were used. We found an amplitude modulation of the late positive slow wave in subtractions, where non-retrieval procedures are mainly used. This amplitude modulation was not evident in additions, where direct retrieval strategies are believed to be used. Our results suggest that the problem size effect is related to non-retrieval procedures of calculation.