Limits...
Memory bias in the temporal bisection point.

Levy JM, Namboodiri VM, Hussain Shuler MG - Front Integr Neurosci (2015)

Bottom Line: The ability to time intervals confers organisms, including humans, with many remarkable capabilities.Finally, by using two sets of referent durations, we showed that only memory bias-corrected measures were consistent with a previously reported effect in which the ratio of the referents affects the location of the bisection point.These results suggest that memory effects should be considered in temporal tasks.

View Article: PubMed Central - PubMed

Affiliation: Department of Neuroscience, Johns Hopkins University Baltimore, MD, USA.

ABSTRACT
The ability to time intervals confers organisms, including humans, with many remarkable capabilities. A common method for studying interval timing is classification, in which a subject must indicate whether a given probe duration is nearer a previously learned short or long reference interval. This task is designed to reveal the probe duration that is equally likely to be labeled as short or long, known as the temporal bisection point. Studies have found that this bisection point is influenced by a variety of factors including the ratio of the target intervals, the spacing of the probe durations, the modalities of the stimuli, the attentional load, and the inter-trial duration. While several of these factors are thought to be mediated by memory effects, the prototypical classification task affords no opportunity to measure these memory effects directly. Here, we present a novel bisection task, termed the "Bisection by Classification and Production" (BiCaP) task, in which classification trials are interleaved with trials in which subjects must produce either the short or long referents or their midpoint. Using this method, we found a significant correlation between the means of the remembered referents and the bisection points for both classification and production trials. We then cross-validated the bisection points for production and classification trials by showing that they were not statistically differentiable. In addition to these population-level effects, we found within-subject evidence for co-variation across a session between the production bisection points and the means of the remembered referents. Finally, by using two sets of referent durations, we showed that only memory bias-corrected measures were consistent with a previously reported effect in which the ratio of the referents affects the location of the bisection point. These results suggest that memory effects should be considered in temporal tasks.

No MeSH data available.


The effect of referent ratio on the bisection point location. Previous work has demonstrated that the bisection point transitions toward the arithmetic mean (AM) from the geometric mean (GM) as the ratio of the referents increases (Wearden and Ferrara, 1996; Allan, 2002b; Kopec and Brody, 2010). Here, we show the cumulative distribution function (CDF) of the distance indexes (defined in Results) for classification and production bisection points. Qualitatively, the data obeys this trend as the median bisection point across all subjects is closer to the AM for the 1/5 s group (referent ratio of 5) than the 2/4 s group (referent ratio of 2) for both classification (left) and production (right) trials assuming (A) a movement time of 0 ms and (B) a movement time of 175 ms. A quantitative treatment of this data can be found in Section “Results.”
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4493391&req=5

Figure 4: The effect of referent ratio on the bisection point location. Previous work has demonstrated that the bisection point transitions toward the arithmetic mean (AM) from the geometric mean (GM) as the ratio of the referents increases (Wearden and Ferrara, 1996; Allan, 2002b; Kopec and Brody, 2010). Here, we show the cumulative distribution function (CDF) of the distance indexes (defined in Results) for classification and production bisection points. Qualitatively, the data obeys this trend as the median bisection point across all subjects is closer to the AM for the 1/5 s group (referent ratio of 5) than the 2/4 s group (referent ratio of 2) for both classification (left) and production (right) trials assuming (A) a movement time of 0 ms and (B) a movement time of 175 ms. A quantitative treatment of this data can be found in Section “Results.”

Mentions: Thus, when D.I = 0, the bisection point equals the bias-corrected GM and when D.I = 1, the bisection point equals the bias-corrected AM. Analyzing the data this way, we found evidence for the referent ratio effect in production trials (but not classification trials for which the power was likely too low). Qualitatively, the mean of the distance index for the 1/5 s group was higher than that of the 2/4 s group for both classification and production trials (Figure 4A), even when assuming a non-zero movement time (Figure 4B). Quantitatively, the production bisection point for the 1/5 s group was significantly different from the GM (Two-tailed Mann–Whitney U-test with n = 10, p = 0.0091, U = 140, z = 2.608; n = 10, p = 0.0046, U = 143, z = 2.8347, for 0 and 175 ms movement times, respectively) but not the AM (p = 1, U = 105, z = 0, for both movement times). While these comparisons were not significant for the 2/4 s group, the production bisection points for the 1/5 and 2/4 s groups were significantly different in closeness to the GM and AM (0.0376, U = 77, z = -2.0788 for both movement times), as expected by the referent ratio effect.


Memory bias in the temporal bisection point.

Levy JM, Namboodiri VM, Hussain Shuler MG - Front Integr Neurosci (2015)

The effect of referent ratio on the bisection point location. Previous work has demonstrated that the bisection point transitions toward the arithmetic mean (AM) from the geometric mean (GM) as the ratio of the referents increases (Wearden and Ferrara, 1996; Allan, 2002b; Kopec and Brody, 2010). Here, we show the cumulative distribution function (CDF) of the distance indexes (defined in Results) for classification and production bisection points. Qualitatively, the data obeys this trend as the median bisection point across all subjects is closer to the AM for the 1/5 s group (referent ratio of 5) than the 2/4 s group (referent ratio of 2) for both classification (left) and production (right) trials assuming (A) a movement time of 0 ms and (B) a movement time of 175 ms. A quantitative treatment of this data can be found in Section “Results.”
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4493391&req=5

Figure 4: The effect of referent ratio on the bisection point location. Previous work has demonstrated that the bisection point transitions toward the arithmetic mean (AM) from the geometric mean (GM) as the ratio of the referents increases (Wearden and Ferrara, 1996; Allan, 2002b; Kopec and Brody, 2010). Here, we show the cumulative distribution function (CDF) of the distance indexes (defined in Results) for classification and production bisection points. Qualitatively, the data obeys this trend as the median bisection point across all subjects is closer to the AM for the 1/5 s group (referent ratio of 5) than the 2/4 s group (referent ratio of 2) for both classification (left) and production (right) trials assuming (A) a movement time of 0 ms and (B) a movement time of 175 ms. A quantitative treatment of this data can be found in Section “Results.”
Mentions: Thus, when D.I = 0, the bisection point equals the bias-corrected GM and when D.I = 1, the bisection point equals the bias-corrected AM. Analyzing the data this way, we found evidence for the referent ratio effect in production trials (but not classification trials for which the power was likely too low). Qualitatively, the mean of the distance index for the 1/5 s group was higher than that of the 2/4 s group for both classification and production trials (Figure 4A), even when assuming a non-zero movement time (Figure 4B). Quantitatively, the production bisection point for the 1/5 s group was significantly different from the GM (Two-tailed Mann–Whitney U-test with n = 10, p = 0.0091, U = 140, z = 2.608; n = 10, p = 0.0046, U = 143, z = 2.8347, for 0 and 175 ms movement times, respectively) but not the AM (p = 1, U = 105, z = 0, for both movement times). While these comparisons were not significant for the 2/4 s group, the production bisection points for the 1/5 and 2/4 s groups were significantly different in closeness to the GM and AM (0.0376, U = 77, z = -2.0788 for both movement times), as expected by the referent ratio effect.

Bottom Line: The ability to time intervals confers organisms, including humans, with many remarkable capabilities.Finally, by using two sets of referent durations, we showed that only memory bias-corrected measures were consistent with a previously reported effect in which the ratio of the referents affects the location of the bisection point.These results suggest that memory effects should be considered in temporal tasks.

View Article: PubMed Central - PubMed

Affiliation: Department of Neuroscience, Johns Hopkins University Baltimore, MD, USA.

ABSTRACT
The ability to time intervals confers organisms, including humans, with many remarkable capabilities. A common method for studying interval timing is classification, in which a subject must indicate whether a given probe duration is nearer a previously learned short or long reference interval. This task is designed to reveal the probe duration that is equally likely to be labeled as short or long, known as the temporal bisection point. Studies have found that this bisection point is influenced by a variety of factors including the ratio of the target intervals, the spacing of the probe durations, the modalities of the stimuli, the attentional load, and the inter-trial duration. While several of these factors are thought to be mediated by memory effects, the prototypical classification task affords no opportunity to measure these memory effects directly. Here, we present a novel bisection task, termed the "Bisection by Classification and Production" (BiCaP) task, in which classification trials are interleaved with trials in which subjects must produce either the short or long referents or their midpoint. Using this method, we found a significant correlation between the means of the remembered referents and the bisection points for both classification and production trials. We then cross-validated the bisection points for production and classification trials by showing that they were not statistically differentiable. In addition to these population-level effects, we found within-subject evidence for co-variation across a session between the production bisection points and the means of the remembered referents. Finally, by using two sets of referent durations, we showed that only memory bias-corrected measures were consistent with a previously reported effect in which the ratio of the referents affects the location of the bisection point. These results suggest that memory effects should be considered in temporal tasks.

No MeSH data available.