Auditory presentation of sentences in hemispheric sentence comprehension

Researcher Eva Farm

Supervisors Dr Jeff Coney

Date: 18th May, 2010


In much of the past research on hemispheric differences in language comprehension, the left hemisphere has been shown to have a significant advantage during sentence comprehension tasks. However more recent research into the area has been starting to show support for an advantage in the right hemisphere during certain kinds of sentence comprehension tasks such as during the comprehension of punch lines to jokes(Coulson & Williams, 2005), and during the comprehension of metaphors(Bottini et al, 1994; Anaki, Faust & Kravetz, 1998).

A model which many of these studies mention is the fine/coarse coding model proposed by Beeman (1998). In this model, Beeman proposes that the left hemisphere employs fine coding during language comprehension, whereas the right hemisphere applies coarse coding. This means that the left hemisphere processes by strongly triggering information that is closely related to the sentence, whereas the right hemisphere operates by more weakly triggering information that is less related to the input. This model is mentioned by studies in order to explain why the left hemisphere shows more facilitation than the right hemisphere when it comes to performing most language tasks, due to the reasoning that the left hemispheres fine semantic coding process allows the left hemisphere to have a fast selection of meanings, allowing it to facilitate quick language processing. Since this is the case, the left hemisphere tends to perform better in these tasks, whereas the right hemisphere is not as efficient due to it employing coarse semantic coding. Since Beeman’s (1998) model states that the right hemisphere activates information in a broader range, it is mentioned in these studies to support, and as an explanation, to why the right hemisphere is found to perform better at understanding jokes and metaphors where the words are not so commonly used and meanings are ambiguous.

The right hemisphere has also been said to be used when creating situation models (Gouldthorp and Coney, 2009b; Ferstl, Rinck & von Cramon, 2005), which are mental representations that people have upon hearing or seeing a description of a situation, which in turn helps explain the presented information to them (Zwaan & Radvansky, 1998). In other words, it is imagining a situation in your head, while someone is describing something to you.

Previous studies have also indicated sex differences in the way males and females utilise their right and left hemispheres. It is quite widely accepted and demonstrated that males tend to be more left lateralised (use their left hemispheres more), whereas females tend to be bilateral (have the ability to use both hemispheres) (Clements et al, 2006; Shaywitz et al., 2000; and Kansaku, Yamaura and Kitazawa, 2001).

Based on these findings from previous research, we decided to investigate the role of the right hemisphere when understanding language through auditory presentation of sentences. It was hypothesised that the right hemisphere would result in more facilitation than the left hemisphere, and that females would be able to perform the task better than the males.


To investigate the hypotheses, participants were asked to listen to one, two, or three sentences through a set of headphones, and were then presented with a related word or a non-word on the computer screen on either the right or left side. These sentences were constructed to provide context to the target word. The more sentences presented, the more context for the target word will be provided. For example, one sentence being presented before the target word ‘PAINT’ would be “I come in different colours.” Two sentences would be “I come in different colours. I am used in art.” While 3 sentences would be “I come in different colours, I am used in art. I can be put on walls.” If a sentence was not meant to prime the coming word, the single sentence presented was “I am a neutral sentence.” The neutral sentence was added in to show a baseline reaction time to show how much facilitation was going on by comparing the other reaction times with the baseline reaction time. If the target word was a non-word, the number of sentences and type of sentence did not affect anything seeing at it would have no priming effect.

If the word appeared on the left side of the screen, the right hemisphere would be used to process it. If it appeared on the right, then the left hemisphere would be used. The combinations of sentences followed by words or non-words were divided into 8 sets of 320 trials. Each set was then randomised 4 times, meaning the order in which they appeared in the program was made random so that each person got a unique set of trials.

Main Results:

Facilitation is shown when you subtract the reaction time of a condition from the mean reaction time for the neutral condition. For example, if we wanted to find out how much facilitation is being demonstrated by the right hemisphere after you have listened to 3 sentences and responded to the target word, then we would take that reaction time, from however long it takes you on average to respond to a target word after you listened to “I am a neutral sentence”.

The results of our study showed no significant difference in facilitation for both hemispheres which did not support our hypothesis. Our hypothesis of females performing better than males was also not supported, however even though the results were not statistically significant, we did notice a trend forming which indicated the that right hemisphere in females seemed to be better than the left hemisphere in all 3 conditions, getting more of a difference between them as the number of sentences increased, with a significant difference after 3 sentences. However more participants are needed in order to find any significant differences or to make any strong claims about sex differences.


We found no differences between the hemispheres in this study which may be explained by the hemispheres possibly applying different processes during the task, which ends up with the same amount of facilitation. It is possible the left hemisphere was applying ‘top-down’ propositional model to predict the word whereas the right hemisphere may have used of a ‘bottom-up’, integrative model.

Although it is recognised that the appropriate statistical evidence is lacking in order to make any strong statements about sex differences in this area, there is an indication from post-hoc analysis of a trend that women may be more responsive to spoken language than men. This opens up an interesting and new direction for research on sex differences in the hemispheres during auditory sentence comprehension including the possibility of sex differences when it comes to using situation models seeing as there was an indication of female right hemisphere superiority.