Semantic processing in aphasia: evidence from semantic priming and semantic interference

ABSTRACT

Semantic processing theories propose activation of concepts via semantic features, with interference from semantic neighbours arising due to shared features. Semantic impairment has been explained as damage to activation and interference mechanisms, and linked to impaired semantic control. This study investigated semantic activation and interference in 20 people with aphasia. We found normal semantic priming or hyper-priming, coupled with significant semantic interference effects, in most of the participants, regardless of scores on standard semantic tasks. There was little evidence of a relationship between executive functions and semantic processing. The data indicate that semantic activation is unimpaired in most people with aphasia. Apparent difficulties with semantic processing are predominantly found when tasks involve resolving competition from close semantic neighbours. These novel findings question the use of offline tasks involving semantic competitors in diagnosis of semantic deficits in aphasia – and other conditions such as dementia – and demand revised diagnostic methods.

KEYWORDS:

• Aphasia
• semantics
• activation
• interference
• priming

Acknowledgements

The authors are extremely grateful to all the people with aphasia and the control participants who contributed to this study. We are also grateful to Richard Body who advised on the selection of tests of executive function. We thank the editor and two anonymous reviewers whose detailed comments enabled us to improve the paper immeasurably.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Author contributions

Lucy Dyson: Conceptualisation, Methodology, Formal Analysis, Investigation, Resources, Data Curation, Writing – Review & Editing, Funding Acquisition. Jane Morgan: Methodology, Formal Analysis, Writing – Review & Editing, Supervision. Ruth Herbert: Conceptualisation, Methodology, Formal Analysis, Resources, Data Curation, Writing original and final version, Visualisation, Supervision, Project administration, Funding acquisition.

Notes

1 As the semantic priming task was central to this study screening for written lexical decision was undertaken to ensure that all participants could complete the task and were responding at rates significantly above chance. Similarly, we aimed to recruit people with a range of semantic impairment hence the need to include a low cut-off score for word to picture matching to ensure those with significant deficits were included but those at floor were excluded.

2 A full description of the test materials is provided in Dyson, L., Morgan, J. & Herbert, R. (submitted). Novel matched stimuli for assessment of lexical semantics. Aphasiology.

3 Filler words were included in order to establish an acceptable relatedness proportion of 0.2 within each word list, after Neely et al. (1989). The relatedness proportion is calculated by dividing the number of semantically similar or associated real word prime-target partners by the total number of test stimuli i.e. 50/250.

4 Websites used for copyright free images include: Morguefile (http://www.morguefile.com/); Flicker Creative Commons (http://www.flickr.com/creativecommons/ http://www.flickr.com/creativecommons/by-2.0/); and Stock.XCHNG (http://www.sxc.hu/). Photographers were contacted regarding the use of their images and will be attributed as per the guidelines of each website in any publication of materials including the images.

5 http://homepages.abdn.ac.uk/j.crawford/pages/dept/Single_Case_Effect_Sizes.htm

6 According to online information from UCLA’s Phonetic Lab: “The highest possible d′ (greatest sensitivity) is 6.93, the effective limit (using .99 and .01) 4.65, typical values are up to 2.0, and 69% correct for both different and same trials corresponds to a d′ of 1.0.” http://phonetics.linguistics.ucla.edu/facilities/statistics/dprime.htm#:~:text=The%20highest%20possible%20d'%20(greatest,corresponds%20to%20a%20d'%20of%201.0.

7 Crawford and Garthwaite (2002) propose that a point estimate of less than 2.5% would be rare in the population and would represent a large deficit in performance. Hence a point estimate that is equal to or larger than 97.5% demonstrates a large advantage in performance (Burgoyne et al., 2016).

8 Three of the participants with aphasia’s accuracy responses in the incongruent condition did not differ significantly from chance and must be viewed with that in mind – participants JC, JK and PS.

9 DB in group 1, and DW in group 3, the anomalous subgroup.

Additional information

Funding

This research was completed as part of a project funded by a Research Fellowship awarded to Lucy Dyson by the Stroke Association [grant number: TSA JRTF 2011/01]