Limits...
Inference by Exclusion in Goffin Cockatoos (Cacatua goffini).

O'Hara M, Auersperg AM, Bugnyar T, Huber L - PLoS ONE (2015)

Bottom Line: Inference by exclusion, the ability to base choices on the systematic exclusion of alternatives, has been studied in many nonhuman species over the past decade.However, the majority of methodologies employed so far are hard to integrate into a comparative framework as they rarely use controls for the effect of neophilia.Our results indicate that Goffin cockatoos are able to solve such abstract two-choice tasks employing inference by exclusion but also highlight the importance of other response strategies.

View Article: PubMed Central - PubMed

Affiliation: Department of Cognitive Biology, University of Vienna, Vienna, Austria; Messerli Research Institute, University of Veterinary Medicine Vienna, Medical University Vienna, University of Vienna, Vienna, Austria.

ABSTRACT
Inference by exclusion, the ability to base choices on the systematic exclusion of alternatives, has been studied in many nonhuman species over the past decade. However, the majority of methodologies employed so far are hard to integrate into a comparative framework as they rarely use controls for the effect of neophilia. Here, we present an improved approach that takes neophilia into account, using an abstract two-choice task on a touch screen, which is equally feasible for a large variety of species. To test this approach we chose Goffin cockatoos (Cacatua goffini), a highly explorative Indonesian parrot species, which have recently been reported to have sophisticated cognitive skills in the technical domain. Our results indicate that Goffin cockatoos are able to solve such abstract two-choice tasks employing inference by exclusion but also highlight the importance of other response strategies.

No MeSH data available.


Related in: MedlinePlus

Percentages response pattern frequencies.a) Detailed patterns exhibited by each individual in all 25 test sessions; values in the bar graphs correspond to the probability that the number of sessions with certain response patterns occurred by chance; significant values are printed bold; b) Frequency of patterns exhibited at group level; horizontal lines indicate median values, boxes span the first to third quartiles and whiskers represent 95% confidence intervals. The dashed horizontal line represents chance levels; the grey area below the line denotes no significant divergence from chance. The letters in the boxplots refer to a Wilcoxon signed-rank test comparing patterns to 'Inference by exclusion'—a: no significance, b: significant, p< 0.05.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4526371&req=5

pone.0134894.g004: Percentages response pattern frequencies.a) Detailed patterns exhibited by each individual in all 25 test sessions; values in the bar graphs correspond to the probability that the number of sessions with certain response patterns occurred by chance; significant values are printed bold; b) Frequency of patterns exhibited at group level; horizontal lines indicate median values, boxes span the first to third quartiles and whiskers represent 95% confidence intervals. The dashed horizontal line represents chance levels; the grey area below the line denotes no significant divergence from chance. The letters in the boxplots refer to a Wilcoxon signed-rank test comparing patterns to 'Inference by exclusion'—a: no significance, b: significant, p< 0.05.

Mentions: Schematic representation of one of each baseline trials, novelty trials and test trials with example stimuli, as well as theoretical response predictions (see main text) colour-coded to match response patterns in Fig 4; + indicates stimuli with positive reward contingency,—indicates unrewarded stimuli; ✓ correct, ✗ incorrect choices. On the right we list the most parsimonious mechanisms (see main text).


Inference by Exclusion in Goffin Cockatoos (Cacatua goffini).

O'Hara M, Auersperg AM, Bugnyar T, Huber L - PLoS ONE (2015)

Percentages response pattern frequencies.a) Detailed patterns exhibited by each individual in all 25 test sessions; values in the bar graphs correspond to the probability that the number of sessions with certain response patterns occurred by chance; significant values are printed bold; b) Frequency of patterns exhibited at group level; horizontal lines indicate median values, boxes span the first to third quartiles and whiskers represent 95% confidence intervals. The dashed horizontal line represents chance levels; the grey area below the line denotes no significant divergence from chance. The letters in the boxplots refer to a Wilcoxon signed-rank test comparing patterns to 'Inference by exclusion'—a: no significance, b: significant, p< 0.05.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4526371&req=5

pone.0134894.g004: Percentages response pattern frequencies.a) Detailed patterns exhibited by each individual in all 25 test sessions; values in the bar graphs correspond to the probability that the number of sessions with certain response patterns occurred by chance; significant values are printed bold; b) Frequency of patterns exhibited at group level; horizontal lines indicate median values, boxes span the first to third quartiles and whiskers represent 95% confidence intervals. The dashed horizontal line represents chance levels; the grey area below the line denotes no significant divergence from chance. The letters in the boxplots refer to a Wilcoxon signed-rank test comparing patterns to 'Inference by exclusion'—a: no significance, b: significant, p< 0.05.
Mentions: Schematic representation of one of each baseline trials, novelty trials and test trials with example stimuli, as well as theoretical response predictions (see main text) colour-coded to match response patterns in Fig 4; + indicates stimuli with positive reward contingency,—indicates unrewarded stimuli; ✓ correct, ✗ incorrect choices. On the right we list the most parsimonious mechanisms (see main text).

Bottom Line: Inference by exclusion, the ability to base choices on the systematic exclusion of alternatives, has been studied in many nonhuman species over the past decade.However, the majority of methodologies employed so far are hard to integrate into a comparative framework as they rarely use controls for the effect of neophilia.Our results indicate that Goffin cockatoos are able to solve such abstract two-choice tasks employing inference by exclusion but also highlight the importance of other response strategies.

View Article: PubMed Central - PubMed

Affiliation: Department of Cognitive Biology, University of Vienna, Vienna, Austria; Messerli Research Institute, University of Veterinary Medicine Vienna, Medical University Vienna, University of Vienna, Vienna, Austria.

ABSTRACT
Inference by exclusion, the ability to base choices on the systematic exclusion of alternatives, has been studied in many nonhuman species over the past decade. However, the majority of methodologies employed so far are hard to integrate into a comparative framework as they rarely use controls for the effect of neophilia. Here, we present an improved approach that takes neophilia into account, using an abstract two-choice task on a touch screen, which is equally feasible for a large variety of species. To test this approach we chose Goffin cockatoos (Cacatua goffini), a highly explorative Indonesian parrot species, which have recently been reported to have sophisticated cognitive skills in the technical domain. Our results indicate that Goffin cockatoos are able to solve such abstract two-choice tasks employing inference by exclusion but also highlight the importance of other response strategies.

No MeSH data available.


Related in: MedlinePlus