Limits...
Arithmetic Memory Is Modality Specific.

Myers T, Szücs D - PLoS ONE (2015)

Bottom Line: In regards to numerical cognition and working memory, it is an open question as to whether numbers are stored into and retrieved from a central abstract representation or from separate notation-specific representations.The participants were presented with numbers (1-9) as either Arabic digits or written number words (Arabic digits and dot matrices in Experiment 2) at the first (S1) and second (S2) stimuli.The participant's task was to add the first two stimuli together and verify whether the answer (S3), presented simultaneously with S2, was correct.

View Article: PubMed Central - PubMed

Affiliation: Centre for Neuroscience in Education, Department of Psychology, University of Cambridge, Cambridge, United Kingdom.

ABSTRACT
In regards to numerical cognition and working memory, it is an open question as to whether numbers are stored into and retrieved from a central abstract representation or from separate notation-specific representations. This study seeks to help answer this by utilizing the numeral modality effect (NME) in three experiments to explore how numbers are processed by the human brain. The participants were presented with numbers (1-9) as either Arabic digits or written number words (Arabic digits and dot matrices in Experiment 2) at the first (S1) and second (S2) stimuli. The participant's task was to add the first two stimuli together and verify whether the answer (S3), presented simultaneously with S2, was correct. We hypothesized that if reaction time (RT) at S2/S3 depends on the modality of S1 then numbers are retrieved from modality specific memory stores. Indeed, RT depended on the modality of S1 whenever S2 was an Arabic digit which argues against the concept of numbers being stored and retrieved from a central, abstract representation.

Show MeSH
S1 Modality x S2 Modality.Vertical bars denote 0.95 confidence intervals for repeated measures ANOVA (Hollands & Jarmasz, 2009). S1 and S2 modality is signified by ‘A’ and ‘D’; ie. Arabic Digit (A) vs. Dot Matrix (D). As an example, AD signifies an Arabic digit at S1 and a dot matrix at S2.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4696806&req=5

pone.0145614.g007: S1 Modality x S2 Modality.Vertical bars denote 0.95 confidence intervals for repeated measures ANOVA (Hollands & Jarmasz, 2009). S1 and S2 modality is signified by ‘A’ and ‘D’; ie. Arabic Digit (A) vs. Dot Matrix (D). As an example, AD signifies an Arabic digit at S1 and a dot matrix at S2.

Mentions: RTs are shown in Fig 7. Data were assessed by a Correctness x S1 modality x S2 modality ANOVA. The .95 confidence intervals were computed for repeated-measures ANOVA (see Hollands and Jarmasz [32]). RTs were 42 ms faster when S1 was an Arabic digit than when it was a Dot Matrix (903 ms vs. 945 ms; S1 Modality: F(1,19) = 7.2177; p<0.0146; Effect size = 0.28; Partial η2 = 0.71). RTs were 350 ms faster when S2 was an Arabic digit (749 ms vs. 1099; S2 Modality: F(1,19) = 485.34; p<0.00001; Effect size = 0.28; Partial η2 = 0.71). Also, as in previous experiments, there was an S1 modality x S2 modality interaction (AA = 696 ms, DA = 802 ms, DD = 1087 ms, AD = 1110 ms; S1 Modality x S2 Modality: F(1,19) = 23.672, p<0.0001; Effect size = 0.55; Partial η2 = 0.99) which showed that, when S2 was an Arabic digit, participants were quicker to verify the sum at S2/S3 when S1 was also a digit. There was no such effect when S2 was a dot matrix. According to S1 modality x S2 modality Tukey post hoc contrasts, there were significant differences between all conditions except between AD and DD (AD vs DD: p<0.2970; AA vs. AD: p<0.0002; AA vs. DA: p<0.0015; AA vs. DD: p<0.0002; AD vs. DA: p<0.0002; DA vs. DD: p<0.0002). There was no interaction between RTs and the correctness of the sum presented at S3 (925 ms vs. 922 ms; Correctness: F(1,19) = 0.0841, p = 0.7; Correctness x S1 Modality x S2 Modality: F(1,19) = 0.8875, p = 0.3).


Arithmetic Memory Is Modality Specific.

Myers T, Szücs D - PLoS ONE (2015)

S1 Modality x S2 Modality.Vertical bars denote 0.95 confidence intervals for repeated measures ANOVA (Hollands & Jarmasz, 2009). S1 and S2 modality is signified by ‘A’ and ‘D’; ie. Arabic Digit (A) vs. Dot Matrix (D). As an example, AD signifies an Arabic digit at S1 and a dot matrix at S2.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4696806&req=5

pone.0145614.g007: S1 Modality x S2 Modality.Vertical bars denote 0.95 confidence intervals for repeated measures ANOVA (Hollands & Jarmasz, 2009). S1 and S2 modality is signified by ‘A’ and ‘D’; ie. Arabic Digit (A) vs. Dot Matrix (D). As an example, AD signifies an Arabic digit at S1 and a dot matrix at S2.
Mentions: RTs are shown in Fig 7. Data were assessed by a Correctness x S1 modality x S2 modality ANOVA. The .95 confidence intervals were computed for repeated-measures ANOVA (see Hollands and Jarmasz [32]). RTs were 42 ms faster when S1 was an Arabic digit than when it was a Dot Matrix (903 ms vs. 945 ms; S1 Modality: F(1,19) = 7.2177; p<0.0146; Effect size = 0.28; Partial η2 = 0.71). RTs were 350 ms faster when S2 was an Arabic digit (749 ms vs. 1099; S2 Modality: F(1,19) = 485.34; p<0.00001; Effect size = 0.28; Partial η2 = 0.71). Also, as in previous experiments, there was an S1 modality x S2 modality interaction (AA = 696 ms, DA = 802 ms, DD = 1087 ms, AD = 1110 ms; S1 Modality x S2 Modality: F(1,19) = 23.672, p<0.0001; Effect size = 0.55; Partial η2 = 0.99) which showed that, when S2 was an Arabic digit, participants were quicker to verify the sum at S2/S3 when S1 was also a digit. There was no such effect when S2 was a dot matrix. According to S1 modality x S2 modality Tukey post hoc contrasts, there were significant differences between all conditions except between AD and DD (AD vs DD: p<0.2970; AA vs. AD: p<0.0002; AA vs. DA: p<0.0015; AA vs. DD: p<0.0002; AD vs. DA: p<0.0002; DA vs. DD: p<0.0002). There was no interaction between RTs and the correctness of the sum presented at S3 (925 ms vs. 922 ms; Correctness: F(1,19) = 0.0841, p = 0.7; Correctness x S1 Modality x S2 Modality: F(1,19) = 0.8875, p = 0.3).

Bottom Line: In regards to numerical cognition and working memory, it is an open question as to whether numbers are stored into and retrieved from a central abstract representation or from separate notation-specific representations.The participants were presented with numbers (1-9) as either Arabic digits or written number words (Arabic digits and dot matrices in Experiment 2) at the first (S1) and second (S2) stimuli.The participant's task was to add the first two stimuli together and verify whether the answer (S3), presented simultaneously with S2, was correct.

View Article: PubMed Central - PubMed

Affiliation: Centre for Neuroscience in Education, Department of Psychology, University of Cambridge, Cambridge, United Kingdom.

ABSTRACT
In regards to numerical cognition and working memory, it is an open question as to whether numbers are stored into and retrieved from a central abstract representation or from separate notation-specific representations. This study seeks to help answer this by utilizing the numeral modality effect (NME) in three experiments to explore how numbers are processed by the human brain. The participants were presented with numbers (1-9) as either Arabic digits or written number words (Arabic digits and dot matrices in Experiment 2) at the first (S1) and second (S2) stimuli. The participant's task was to add the first two stimuli together and verify whether the answer (S3), presented simultaneously with S2, was correct. We hypothesized that if reaction time (RT) at S2/S3 depends on the modality of S1 then numbers are retrieved from modality specific memory stores. Indeed, RT depended on the modality of S1 whenever S2 was an Arabic digit which argues against the concept of numbers being stored and retrieved from a central, abstract representation.

Show MeSH