The aim of the present fMRI study was to investigate the influence of the verbal-visual cognitive style on cerebral activation patterns during mental calculation. The habitual use of visualization and verbalization during mental arithmetic of each of the 42 right-handed participants was assessed with a short self-report measure. In the first functional session, subtraction and multiplication problems were presented, and in the second functional session, multiplications were presented in two formats, either as Arabic numerals or as written number words.
With regard to verbalization, we found that the higher the self reported tendency to verbalize during mental arithmetic the higher the activation in brain areas related to language or auditory processing, namely, within the right and left supramarginal gyrus, the right and left Rolandic operculum, and the right and left Heschl's gyrus. The supramarginal gyrus is a region in the inferior parietal lobe, and has been found to be involved in phonological processing [28
], reading both in regards to meaning [29
] and phonology [30
], word production [31
], and grammar learning [32
]. The Rolandic operculum has been found to be a somatosensory region [33
], involved in auditory processing, activated by listening to the sound of one's own voice [35
], and the processing of prosody [36
]. The left Rolandic operculum is assumed to be involved in syntactic encoding during speaking [38
], and phonological rehearsal [39
]. The right Rolandic operculum has been associated with the processing of sentence intonation [40
], and of slow prosodic modulations [41
]. The activation of the right Rolandic operculum increased with the degree an individual relied on verbalization during mental arithmetic when the problems were presented with written number words. This result suggests that verbalizers imagine hearing the sound of their voices mentally while reading numbers presented with number words, confirming their own subjective self-report. The right and left Heschl's gyrus is found in the area of the primary auditory cortex in the superior temporal gyrus of the human brain, the first cortical structure to process incoming auditory information [42
]. Naming numbers has been shown to be dependent on linguistic properties. For example, naming latencies for two-digit numbers increase with syllable length [43
]. The authors suggested that digits are translated into a verbal code before being processed [43
It is unclear, why we found significant correlations only for the verbalizer dimension but not for the visualizer dimension. One reason might be that participants are not very good at self-reporting their habits in information processing. We think it would be a major contribution for further studies in this field to design a more valid questionnaire about cognitive styles in mental arithmetic. A limitation of the present study might be that the self-reported use of visualization in mental arithmetic was limited to the visualization of numbers and (intermediate) results while calculating. Other types of visualizations, such as to move mentally along the number line, imagining a mass that increases or decrease in magnitude, or some collection of dots that grows or shrinks in number and others were not included. Another reason may be that the statistical spread of the self-reported use was greater for the verbalizer dimension than for the visualizer dimension (Var(score verb = 1.15) > Var(score vis = .87) ).
It is surprising that the correlations for the first functional session between individual differences in verbalization and brain activation during calculation were only found for the subtraction but not for the multiplication problems (see Figure ). It has frequently been argued that multiplication relies more on verbal processing than subtraction, which is assumed to rely more on visuo-spatial processing [2
]. Therefore, we would have expected verbalizers to show greater language-related activation for multiplications. One possibility why we did not find significant activation patterns for the multiplication problems in the first functional session could be that the multiplication problems used here were very easy to solve. They were part of the multiplication tables and the results could be retrieved from long-term memory. The behavioral results showed that the presented subtraction problems were more difficult than the presented multiplication problems. It is possible that differences with regard to verbalization as cognitive style emerge only for more difficult arithmetic problems that have to be solved in several processing steps leading to intermediate results that have to be kept in memory.
Figure 7 ROI analysis results: Multiplication problems in the first functional session. Correlations between peak activations in the left supramaginal gyrus, the left Rolandic Operculum, and the left Heschl's gyrus and the self-reported use of verbalization: In (more ...)
With regard to a possible domain-specificity (IPS) or modality-specificity (AG) of brain areas involved in number processing we did not observe any modulation. Although we found significant brain activation as a function of verbalization in areas related to sound and language processing, we observed no modulation of activation in the left AG as a function of the self-reported tendency to use verbalization while performing mental arithmetic. The left AG is assumed to support the long term memory retrieval for arithmetic fact knowledge [20
]. Arithmetic fact knowledge is required, for example, in the skilled solving of multiplication problems by retrieving the result from the verbal long-term memory, namely, from the multiplication tables learned in childhood. The left AG shows stronger activation for solving arithmetic problems for which participants report fact retrieval whereas the application of procedural strategies is accompanied by widespread activation in a fronto-parietal network [46
]. These findings link the left AG to arithmetic fact retrieval. In the present study, the activation of the left AG was not modulated by verbalization, and it could be concluded that the left AG shows no specific affinity to the verbal domain and subserves number processing in a modality-general way. This interpretation corresponds to some recent findings [21
] which also raise some doubts about the assumption that the left AG mediates verbal fact retrieval during multiplication.
In the first functional session, we found that subtraction problems activated the right and left IPS more strongly than multiplication problems, whereas multiplication problems activated the left and right angular gyrus more strongly than subtraction problems. This finding corresponds well to results reported in previous studies [47
] as well as to the model of Dehaene and collegues [20
]. It should be noted, however, that multiplication problems were solved faster and more accurately than subtraction problems, although problem sizes were identical. It can therefore not be excluded that some of the observed activation differences are also due to a difference in task difficulty.
In the second functional session, we found that multiplications presented with written number words activated areas implicated in visual processing more strongly than multiplications presented with Arabic numerals. It is highly likely that the greater activation in visual areas for multiplications presented with written number words is due to the larger number of characters. The multiplications in written number words were presented using three words with several letters, whereas the multiplications presented with Arabic numerals consisted of two Arabic numerals and a multiplication sign. Other studies also reported format effects in arithmetic. It has been suggested that numbers presented in different surface-formats have differential access to number representations [49
]. Format differences were also observed in an EEG-study, with more negative event-related potentials for written number words than for Arabic numerals and auditorily presented number words [50
]. The modality-dependent access to numerical information may be a consequence of modality-dependent access to the number representation in parietal cortex. A transcranical magnetic stimulation experiment showed a dissociation between digits and number words in the right parietal lobe, whereas the left parietal lobe showed a double dissociation between the different numerical formats [51
]. Typically, problems presented with number words take longer to solve than problems presented with Arabic numerals [52
]. In our study, multiplications presented with Arabic numerals were also solved faster than multiplication problems presented with written number words.
The visual-verbal cognitive style is assumed to be a relatively stable characteristic although it might depend on the task [5
]. An individual might, for example, prefer visualization for solving arithmetic problems and verbalization for memorizing a poem. It is therefore preferable to assess the visual-verbal cognitive style specifically for the tested task domain. We tried to assess the visual-verbal cognitive style during mental arithmetic with a short self-report measure. Our results indicate that people who say they verbalize more show more activity in brain areas related to language and auditory processing. It is unclear, however, in how far these results are specific to mental arithmetic.
A cognitive style is assumed to be a relatively stable characteristic that describes an individual's way to process information [4
]. Consequently a questionnaire that assesses the habitual use of visualization and verbalization has to measure an independent construct and not personality or intelligence [54
]. Previous studies already showed independence of the visual-verbal cognitive style from intelligence and personality [56
]. Regarding intelligence, we found no significant correlations between the verbal or visual cognitive style and verbal, numerical or figural intelligence. This indicates that the correlations between cognitive style and brain activation are not due to differences in intelligence.
The aim and results of the present study might be understood by some researchers as advocating a view that different learning styles have real consequences for the brain and that education should be adapted accordingly. It is important to distinguish between cognitive styles and learning styles. Cognitive styles are assumed to be an individual's way to process information, whereas learning styles are concerned with the learning environment [58
]. Some studies showed correlations between cognitive styles and some environmental conditions, especially the preferred learning mode. There is a tendency for visualizers to use pictures and for verbalizers to prefer writing as working mode or learning mode [59
]. In our study, the self-reported cognitive style influenced activity in brain areas directly related to the preferred modality of information processing.