Variation in semantic priming across age groups: An AERP study

(1)

The development of word processing skills in children, and its subsequent decline in elderly persons both involve some degree of semantic processing; it is useful to ask, therefore, how semantic processing changes across the lifespan. There is, however, a scarcity of data about how developmental and age-related declines infl uence semantic processing. The goal of this research was to study and compare children, young adults, and seniors, as they engaged in a semantic processing task, indirectly involving priming of a word by a preceding word.

Auditory event-related potentials (AERPs) recorded by scalp electrodes have demonstrated sensitivity to semantic categorization (Kutas & Hillyard, 1984), extending our understanding of the neurobiology underlying this process. The auditory ERP waveform generated by a word consists of a series of positive and negative voltage defl ections or components in the electroencephalographic (EEG) activity evoked by the word. Figure 1 illustrates three components typically observed; an initial negativity, the N1 component, peaking in the 100-ms latency region, followed by a positive peak (P2) at a latency in the region of 200 ms, followed by a prolonged

negativity extending over the latency range from about 400 to about 1000 ms. This prolonged negativity is best described as “ processing negativity ” or PN. Previous research (Mehta et al, 2009) has demonstrated that it refl ects at least four processes: attention, pho- nological processing, semantic processing, and immediate memory.

The PN component in word processing has often been labeled the

“ N400 ” because of the pioneering study of Kutas and Hillyard (1980). In studying the effect of replacing the last word of a meaningful sentence with another word, incongruous with the sense of the sentence, they discovered what came to be called the “ N400 effect ” , an augmentation of the processing negativity evoked by the incongruous word relative to the congruous word. Since this difference peaked at a latency of about 400 ms, the N400 label came to be attached to any negativity following the P2 peak, no matter how it had been generated. Subsequent research has demonstrated, however, that the negativity appears whenever a word is heard in the context of a meaningful auditory task, and that it includes more than one component overlapping in time, refl ecting the simulta- neous parallel processing of more than one cognitive dimension.

Original Article

Variation in semantic priming across age groups: An AERP study

Jyutika Mehta * & James Jerger

^†

* Department of Communication Sciences and Disorders, Texas Woman ’ s University, Denton, USA, and ^† School of Behavioral and Brain Sciences, The University of Texas at Dallas, USA

Correspondence: Jyutika Mehta, Department of Communication Sciences and Disorders, Texas Woman ’ s University, P.O. Box 425737, Denton, TX 76204-5737, USA.

E-mail: [email protected]

(Received 8 August 2013 ; accepted 13 December 2013 )

Abstract

Objective: To study the semantic priming effect on words across the life span by means of auditory event-related potentials (AERPs).

Design: Participants heard a series of three words (S1, S2, and S3). The task was to indicate whether S2 was in the same semantic category as S3. Semantic priming was quantifi ed as the difference between AERPs to the second word when it was semantically related to the fi rst word (S2-R) or unrelated to the fi rst word (S2-UR). Interest was focused entirely on the processing negativity (PN) component of the AERP to S2. The purpose of S3 was to delay the task decision so that the LPC generated by the decision would not confound the measurement of the PN component to S2. Sample: Ten children (9 – 11 years), 11 young adults (20 – 30 years), and 10 seniors (60 – 70 years). Results: The semantic priming effect was evident in the difference between peak amplitude of the PN component of the AERP to S2R and S2UR in all three groups. Children showed a clear asymmetry favoring the left hemisphere. In young adults, the asymmetry still favored the left hemisphere, but the degree of asymmetry was less robust. In the case of seniors, the priming effect was greater over the right hemisphere. Conclusion: Results suggest that all age groups benefi t from contextual support, as evidenced by the semantic priming effect. However, differences in hemispheric asymmetry of activation indicate that perhaps seniors may need to recruit additional, but somewhat different brain resources to manage otherwise largely automatic tasks.

Key Words:

Semantic priming ; auditory event-related potential ; processing negativity ; age groups ; hemispheric asymmetry

(2)

In this paper we use the terminology PN rather than N400 in referring to this electrophysiological manifestation of the complex processing of heard words.

When two consecutive words are related, the amplitude of the PN wave elicited by the second word is attenuated when compared with the condition in which the words are unrelated (Deacon et al, 1995). In semantic priming tasks the amplitude of the PN waveform of the ERP elicited by a word that is related to a preceding or prime word is compared to the amplitude of the PN waveform of the ERP elicited by an unrelated word. The PN component elicited by the related word is less negative than the PN component elicited by the unrelated word. This is the semantic priming effect ; it is illustrated in Figure 2 as the difference between the waveforms elicited by the related preceding word versus the unrelated preceding word.

It is generally believed that, in contrast to age-related declines in lexical retrieval processes, semantic priming effects are relatively unaffected by normal aging (Laver & Burke, 1993). However, there is some debate about whether the size of the semantic priming effect remains stable across the adult lifespan (Rabinowitz, 1986; Mueller et al, 1997). On one hand, studies have shown that the structure of the semantic network is not qualitatively affected by normal aging (Laver & Burke, 1993; Madden, 1988), on the other hand, studies also report that the speed with which semantic information is accessed and integrated changes with age (Birren & Fisher, 1995;

Kutas & Iragui, 1998; Millberg & Blumstein, 1981). To date, debate still continues as to whether this is a consequence of non-specifi c cognitive and motoric decline or specifi c to semantic processing.

Although there have been several studies on semantic priming in

adults (Grose-Fifer & Deacon, 2004; Kutas & Iragui, 1998) as well as children (Gray et al, 1999; Markson & Bloom, 1997; Rice et al, 1992), to our knowledge, none of them have investigated these effects in children, young adults, and seniors in the same paradigm.

The discrepancies in various studies (some reporting signifi cant developmental and age-related changes and some others showing no differences) could possibly arise as a result of diverse factors such as: (1) inconsistent linguistic content (use of nouns and verbs) in stimuli, (2) varied duration of the interval between the prime (S1) and the target (S2) words, (3) variable proportion of related to unrelated words, (4) differences in the strength of relationship between the prime (S1) and the target (S2) words, (5) inconsistent use of words and non-words in the same paradigm, and (6) use of words/non-words of varied length and syllables.

The ERP methodology has been successfully applied in research to investigate word and language processing (Kutas & Hillyard, 1980; Molfese, 1995; Kok, 2001) over past decades. Some AERP studies have investigated semantic processing in children and adults (Juottonen et al, 1996; Holcomb, 1993; Atchley et al, 2006) using mainly sentences as stimuli. In general these studies have investigated a later cognitive potential (P300), the late positive component (LPC), which is dependent on task-relevant classifi cation of stimuli and the decision-making process rather than on the actual semantic processing of the words. LPC amplitudes and latencies are sensitive to the cognitive effort, time required to evaluate and classify stimuli, and the relative diffi culty with which an appropriate decision is reached (Picton, 2011). Also, they may not discriminate processing ability from a variety of other developmental or age-related issues which may lead to LPC differences.

Hemispheric asymmetry

Both the left and right hemispheres make important contributions to the process of language comprehension. A few ERP studies have revealed hemispheric differences in semantic processing, mainly the way in which context infl uences word recognition or semantic priming (Chiarello et al, 1990; Chirarello & Richards, 1990). These studies investigated hemispheric activation to words. They found Abbreviations

AERP Auditory event-related potential EEG Electroencephalographic LPC Late positive component PN Processing negativity S2R S2 semantically related to S1 S2UR S2 unrelated to S1

Figure 1. AERP waveform illustrating three components observed during word processing. There is an initial negative peak (N1) at about 100 ms latency, followed by a positive peak (P2), in the 200-ms latency region. Processing negativity (PN) follows P2 over the latency range from about 400 to 1000 msec.

Figure 2. AERP waveforms illustrating the semantic priming effect.

Two words are presented successively under two conditions. In one condition, the fi rst word is semantically related to the second word.

In the other condition, the two words are unrelated. The amplitude of the PN component of the waveform for the ‘ related word ’ condition is less negative than for the ‘ unrelated word ’ condition. This is the ‘ priming effect ” . It extends over the latency range from about 400 ms to 1000 ms.

(3)

that there is a rapid selection of one meaning and suppression of other candidates in the left hemisphere, while activation spreads more diffusely in the right hemisphere in young adults. However even fewer studies have examined AERP hemispheric differences to semantic processing in different age groups. Holcomb et al (1992) showed that, in individuals from 5 to 26 years, the scalp topography of PN varied considerably. Children showed a relatively more left and anterior distribution of the PN waveform than in young adults.

This asymmetry perhaps refl ects maturational changes in semantic processing abilities as the individual develops more facility with advancing age.

Neuroimaging research has shown that older adults recruit or acti- vate more brain areas while performing cognitive tasks (Cabeza, 2001; Cabeza et al, 2004; Kahlaoui et al, 2007). These studies suggest an age-related reduction of hemispheric asymmetries while performing memory and language-related tasks. Seniors exhibited a bilateral hemispheric activation, while younger adults showed left- hemispheric activation for language based tasks, suggesting that substrates for the semantic processing of words change as we age.

On the other hand, Johnson and colleagues (2001) showed that there were no signifi cant differences between younger and older adults in fMRI activation during a semantic categorization task after account- ing for differences in motor and processing speed. Understanding how semantic processing changes with normal development and aging may provide insights into processes that may get disrupted as a result of neurodegenerative conditions.

The goal of the present study was to determine how the peak amplitude of the PN component and its hemispheric topography

changes across age groups. To this aim we employed a modifi ed version of the classical semantic priming paradigm, in which two sequentially presented words, which are either related or unrelated, were followed by a third word. The participant ’ s task was to indicate whether or not the third word was semantically related to either of the two words that preceded it.

The purpose of the three-word trial design was to defer the decision about the second word until the third word had occurred. In ERP component analyses, often the PN is contaminated by the subsequent LPC. Figure 3 illustrates how LPC and PN interact and infl uence each other. In order to ensure that the late positive component (LPC) elicited by the decision-making process did not contaminate the PN associated with the second word; the latter was expected to be smaller in amplitude when it was preceded by a related word and relatively larger when it was preceded by an unrelated word. The difference between the PN components of related and unrelated second words was interpreted as an ERP index of the semantic priming effect.

Methods

The entire study protocol was approved by the Institutional Review Board (IRB) of the University of Texas at Dallas.

Participants

Thirty-one volunteer participants were recruited in three groups: ten in the children ’ s group (ages 9 – 11 years; 5 males and 5 females),

Figure 3. Progressive changes in morphologies of LPC and PN components of the AERP to words across posterior to anterior midline electrode sites. The solid curves are grand averages from target data; the dashed curves are non-target responses. At electrode OZ, the LPC, refl ected in the solid curve, is prominent, but the PN is also clearly visible in both waveforms. As the electrode site moves forward along the midline, progressively increasing negativity of the PN component reduces, and ultimately eliminates the LPC component in the target waveform. There is a continuous interaction between the two conditions.

(4)

eleven in the young adult group (20 – 30 years; 6 males and 5 females), and ten in the senior group (60 – 70 years; 4 males and 6 females). All participants were right-handed, native speakers of American English, and had no history of a neurological, psychiatric, medical, or speech/

language disorder.

Initial audiometric measures

Each participant was evaluated by standard pure-tone and immittance audiometry in order to ensure that peripheral hearing sensitivity was within the normal range. A pure-tone audiogram was obtained at octave intervals from 250 to 6000 Hz on a two-channel, clinical audiometer (Grason-Stadler, model 16). Potential participants were seated in a medical examining chair and fi tted with standard audiometric earphones. Participants responded to tone presentations by pressing a response button. Thresholds were defi ned by adherence to the Hughson-Westlake protocol for threshold searches. A potential participant was excluded from the study if any pure-tone threshold exceeded 25 dB HL on either ear over the frequency range from 250 – 2000 Hz, or exceeded 35 dB HL at 4000 or 6000 Hz. On immittance measures all participants showed type A tympanograms and ipsilateral stapedial refl exes at 95 dB HL on both ears.

In order to further ensure that audibility of the experimental stimuli would not pose an issue in data interpretation, all experimental stimuli were presented at the individual participant ’ s self-reported most comfortable loudness level. This level was determined by seating the participant directly in front of the loudspeaker used to present subsequent experimental stimuli, and presenting the short sentence “ where is it coming from ” . Beginning at a faint level the intensity of the sentence was increased in 2-dB steps until the participant judged it “ too loud ” ; then it was decreased in 2-dB steps until the participant judged it “ most comfortable ” . This level was noted for each participant; all subsequent experimental words were presented at this pre-determined presentation level.

The loudspeaker was at the participant ’ s ear level, at a distance of 2.2 metres.

Stimuli

A total of 480 words were recorded in a sound-treated room by an adult male, monolingual speaker of English. Speech was sampled at a rate of 22 050 Hz with 16-bit amplitude resolution using Cool Edit Pro ^TM 2.1 (Syntrillium Software Corporation, 2003). The mean duration of all words was calculated to be 550 ms (SD ⫽ 12 ms).

The selected words were category exemplars from the Battig &

Montague (1969) category norms. All stimulus words were mono- syllabic, concrete, imageable, familiar nouns (Coltheart, 1981) with low age of acquisition scores (Gilhooly & Logie, 1980). Overall intensities of the words were digitally equated based on their average root mean square power.

As discussed earlier, the purpose of the three-word trial design was to defer the decision about the second word until another word had occurred in order to ensure that the late positive component (LPC) elicited by the decision-making process did not contaminate the PN associated with the second word, these words were presented in triplets. The following conditions applied to each triplet set:

The fi rst word (S1) of each triplet was always unrelated to the 1.

third word (S3) of the preceding triplet.

The second word (S2) was related to S1 in 50% of the trials 2.

(S2-R) and unrelated in the other 50% (S2-UR).

The third word (S3) of each triplet was related to the S1 in 12.5%

3.

of the trials, to the S2 in 12.5% of the trials, to both words in 25% of the trials, and to neither word in 50% of the trials.

There were a total of 160 word triplets with 80 in the S2-R condition and 80 in the S2-UR condition.

Two examples of word triplets are as follows:

fork, knife, dog (S2-related to S1) 1.

sun, phone, moon (S2-unrelated to S1) 2.

The target stimuli were delivered to the participants by a loudspeaker located directly in front of the participant at a distance of 2.2 metres, via the Neuroscan Stim ² presentation software (Stim ² , Compumedics).

ERP recording

Continuous electroencephalographic (EEG) activity was recorded from 30 scalp locations using silver-silver chloride electrodes attached to an elastic cap (Neurosoft) and a Neuroscan electrophysiological data acquisition system (SCAN 4.2, Compumedics Ltd.).

All electrode impedances remained below 10 k Ω . Two electrodes placed above and at the outer canthus of the left eye monitored eye movement and blinks. EEG channels were referenced to linked mastoid electrodes with Fpz as ground. Ongoing EEG activity was sampled at 1000 Hz, amplifi ed, analog-fi ltered from 0.15 to 70 Hz, digitized, and stored for later off-line analysis. A trigger initiated each epoch at the onset of each word.

Experimental task

Each trial consisted of three words presented sequentially: S1, S2, and then S3. An interstimulus interval (ISI) of 2200 ms (from offset of one word to onset of the next word) separated S1 and S2, and S2 and S3. An intertrial interval (ITI) of 5 s was interposed between S3 of one triplet and S1 of the next. Participants were asked to listen to all the words of the triplet but to respond only to the third word (S3) of each triplet. The task was to indicate whether S3 was semantically related to either S1 or S2, or both S1 and S2 ( ‘ yes ’ response), or to neither ( ‘ no ’ response). A pad with two buttons, one button for the ‘ yes ’ response and one for the ‘ no ’ response, was provided with the buttons placed side-by-side vertically and separated by a distance of 1.2 inches. The participants were instructed to use only the index fi nger of the right hand to respond.

Data analysis

Off-line, individual epochs, encompassing ⫺ 200 to 1600 msec relative to stimulus onset, were derived for each participant.

Epochs were rejected if activity in the eye channel exceeded ⫾ 50 microvolts. Following artifact rejection, epochs were separately averaged, linearly detrended, and digitally low-passed fi ltered at 20 Hz (fi lter slope ⫺ 48 dB/octave) then averaged for each condition. The averaged waveforms of individual participants were the grand averaged.

The semantic priming effect was quantifi ed as the difference in the mean peak amplitudes of word S2-R and word S2-UR over the latency range from 250 – 750 msec (region of processing negativity).

These procedures were repeated for determining mean peak amplitudes for all S1 words. The S2 peak-amplitude data were statistically analysed using a repeated measures analysis-of-variance (ANOVA, StatView, SAS Institute, Inc., 1998). Group affi liation served as the

(5)

between-subjects factor; electrode site and condition served as the within-subject factors. Statistical signifi cance was evaluated at the 0.05 alpha error level.

Results

The children and young adults showed the largest peak amplitudes over the left hemisphere electrodes, specifi cally over electrode CP3, while the seniors showed the largest effect over the right hemisphere electrodes, specifi cally over electrode CP4. For this reason we chose the CP array of electrodes over the left and right hemispheres, electrodes TP7, CP3, CP4, and TP8, for further analysis.

Figure 4 shows grand-averaged waveforms for S2 words across the three age groups. Within each group the priming effect is illustrated by comparison between the AERP waveform for S2-R words versus the waveform for S2-UR words. The absolute amplitudes of both PN components were largest in the children, smallest in the seniors. Within each age group, moreover, the depth of the PN component was greater for S2-UR words than for S2-R words,

the classical priming effect. It is possible, however, that an unsus- pected difference between the two conditions may account for the difference in S2 waveforms. Arguing against this possibility are the waveforms for S1 words, illustrated in Figure 5. Since the S1 word occurred before the S2 word, it should not have mattered to the S1 word whether or not the S2 word was related to it. That was indeed the case. Figure 5 shows that the peak amplitude of the PN components for the S1 word, categorized according to relationship to the S2 word, were similar.

A repeated measures analysis of variance (ANOVA) was carried out on the PN peak amplitude measures of S2 words.

There were three main effects, age group (children, young adults, seniors), condition (S2-R versus S2-UR), and electrode (TP7, CP3, CP4, and TP8). In order to highlight the hemispheric asymmetries, midline electrode CPZ was omitted from this analysis. Table 1 shows the results of the ANOVA. All three main effects were signifi cant; Group, p ⫽ 0.018, Condition, p ⫽ 0.002, Electrode, p ⫽ 0.0002. In addition, there was a signifi cant Electrode ⫻ Group interaction, and a signifi cant Figure 4. Grand averaged waveforms of S2 Words. Comparison of S2-related to S1 (S2R) and S2-unrelated to S1 (S2-UR) conditions, revealed signifi cant differences between the two conditions in all three participant groups. The differences in the PN in these waveforms indexed the semantic priming effect in all three groups. The maximal difference was observed at electrode CP3 in children and young adults, and at electrode CP4 in seniors. Comparison is between the S2-R and S2-UR conditions.

Figure 5. Grand averaged waveforms of S1 words. Comparisons of S1-R and S1-UR conditions show maximal response at electrode CP3 for children and young adults, and at electrode CP4 for the seniors. There were no obvious differences between the two conditions in any participant group.

(6)

Condition ⫻ Electrode interaction, p ⫽ 0.0044. The Condition ⫻ Group interaction, however, was not signifi cant, p ⫽ 0.43.

The signifi cant group effect results largely from the variation in absolute peak amplitude across the three age groups. Such variation is not unexpected, especially when children and elderly persons are involved. As shown below, when all data have been transformed to equate them for absolute amplitude, i.e. to remove the signifi cant variation in absolute amplitude across groups, the data can be more equitably viewed. In the original ANOVA, however, the appropriate test of whether the semantic priming effect varies signifi cantly across age groups is the Condition ⫻ Group interaction.

As noted above, in order to illustrate these results graphically it was fi rst necessary to equate the three groups for overall absolute amplitude. To this end we transformed the mean amplitude condition differences (S2-UR versus S2-R) for each group to an equivalent scale according to the following formula:

TA (transformed amplitude) ⫽ (S2-UR) ⫺ (S2-R)/S2-UR ⫹ S2-R) These transformed mean amplitude differences are illustrated in Figure 6. The data at electrode CPZ are included for better

visualization of the interactions. The priming effect was clearly greater over the left hemisphere in the children, less so in young adults, and greater over the right hemisphere in the senior group.

To illustrate this systematic shift in hemispheric asymmetry graphically we calculated a hemispheric asymmetry index (HAI) from the mean transformed peak amplitudes at the four lateral electrode sites on the CP array as follows:

HAI ⫽ (TP7 ⫹ CP3)/(CP4 ⫹ TP8)

Note that HAI indices greater than 1.0 indicate asymmetry toward the left hemisphere, while indices less than 1.0 refl ect greater activity over the right hemisphere. The present results are plotted for each group in Figure 7. The HAI index was strongly asymmetric to the left hemisphere in children (2.0), less so in young adults (1.57), and was shifted to the right hemisphere in the seniors (0.56).

Discussion

Mean peak amplitude differences between S2-R and S2-UR for the PN component indexed the semantic priming effect. The mean peak amplitude for the PN component of S2-R was signifi cantly smaller than the mean peak amplitude for the PN component of S2-UR for all three participant groups. The fact that the Condition ⫻ Group interaction was not signifi cant leads to the conclusion that the priming effect was essentially equivalent across the age span.

This is consistent with the conclusion previously reached by Laver & Burke (1993).

While the transformed measures of the priming effect suggest little change across groups when averaged across electrodes, the pattern across the CP array indicates that hemispheric asymmetry shifts systematically from strong left hemisphere asymmetry in the Children Group, to less strong right hemisphere dominance in the Senior Group (Figure 7). We discuss this further in the next section.

Hemispheric differences

A signifi cant interaction between Group and Electrode array confi rmed that hemispheric asymmetry changed across the lifespan.

Figure 6. The semantic priming effect in the transformed peak amplitude measures across the centro-parietal array of electrodes. The priming effect was greatest over the left hemisphere (electrode CP3) for children and young adults but greatest over the right hemisphere (electrode CP4) for seniors. The Group ⫻ Electrode interaction was signifi cant, confi rming a shift in hemispheric asymmetry across the groups in the semantic priming effect.

Table 1. Analysis of variance for S2 peak amplitude.

Group ⫽ children, young adults, seniors. Condition ⫽ S2 R, S2 UR.

Electrode ⫽ TP7, CP3, CP4, TP8.

Variable Mean square DF F p

Group 90.67 2 4.63 0.018

Subject (Group) 19.59 28

Condition 39.47 1 12.04 0.002

Condition ⫻ Group 2.85 2 0.87 0.43

Condition ⫻ Subject (Group) 3.28 28

Electrode 32.04 3 7.58 0.0002

Electrode ⫻ Group 14.07 6 3.71 0.0007

Electrode ⫻ Subject (Group) 4.230 84

Condition ⫻ Electrode 6.51 3 4.708 0.0044

Condition ⫻ Electrode ⫻ Group 1.77 6 1.28 0.276

Condition ⫻ Electrode ⫻ Subject 1.38 84

Figure 7. Systematic shift in hemispheric asymmetry index (HAI) across the three age groups. When HAI ⫽ 1.0 there is balanced hemispheric symmetry. Indices greater than HAI ⫽ 1.0 indicate asymmetry favoring the left hemisphere; indices less than HAI ⫽ 1.0 indicate asymmetry favoring the right hemisphere.

(7)

Children and young adults exhibited the maximal effect at electrode CP3 (left hemisphere), while seniors exhibited the maximal effect at CP4 (right hemisphere). Overall, the hemispheric representation of the semantic priming effect was most asymmetric toward the left hemisphere for children and for young adults, and asymmetric toward the right hemisphere for seniors. Past neuroimaging studies support these results (Cabeza, 2001; Cabeza et al, 2004). Seniors exhibit increased activation of brain areas while performing certain cognitive tasks.

Cabeza (2001) suggests that the brains of seniors show age-related reduction of functional asymmetries when compared with younger adults. Although this pattern has been consistently shown for memory and perception processes (Velanova et al, 2001), there is also some evidence for reduction of hemispheric asymmetries for semantic processing (Kahlaoui et al, 2007). Kahlaoui et al (2007) suggest that the recruitment of the right hemisphere for semantic processing in seniors is a necessary compensatory mechanism to preserve language abilities in successful aging. Overall age-related declines include shrinkage of brain tissue (Haug & Egers, 1991), lesser concentration of neurotrans- mitters (Woodruff-Pak, 1997), and reduction in hemodynamic factors refl ecting neurocognitive declines (D ’ Esposito et al, 1999). All of these factors may lead to general defi cits in the processing of semantic information: seniors may have to recruit additional areas of the brain to maintain a similar level of performance. This may be the factor underlying the reversed asymmetry observed in our senior group.

In summary, the fi ndings of the present study suggest that the priming effect, per se, does not change substantially across the age span, but that hemispheric asymmetry of the semantic priming effect changes across age groups. There have been few studies examining children, young adults, and seniors in the same paradigm, and using the same stimuli. Since the present study allowed examination of the semantic priming effect across different age groups, using the same word stimuli, the fi ndings speak to the semantic priming changes that occur with age. The current data suggests that everyone benefi ts from contextual support, as evidenced by the equivalent semantic priming effect in all age groups. However, age-related differences in hemispheric recruitment indicate that perhaps seniors may need to recruit additional, but somewhat different brain resources to manage otherwise largely automatic tasks.

Acknowledgements

We are grateful to the participants and the two reviewers for helpful criticism. The experimental protocol was approved by the Institu- tional Review Board of the University of Texas at Dallas.

Declaration of interest: The authors report no confl icts of interest.

The authors alone are responsible for the content and writing of the article.

References

Annett M . 1970 . Classifi cation of hand preference by association analysis . Br J Psychol , 61 , 303 – 321 .

Atchley R. , Rice M. , Betz S. , Kwasny K. , Sereno J. et al . 2006 . A comparison of semantic and syntactic event related potentials generated by children and adults . Brain Lang , 99 , 236 – 246 .

Atchley R. , Story J. & Buchanan L . 2001 . Exploring the contribution of the cerebral hemispheres to language comprehension defi cits in adults with developmental language disorder . Brain Cogn , 46 , 16 – 94 .

Battig W. & Montague W . 1969 . Category norms for verbal items in 56 categories . A replication and extension of Connecticut norms. Journal of Experimental Psychology Monographs , 80 , 1 – 45 .

Bentin S. , Kutas M. & Hillyard S . 1993 . Electrophysiological evidence for task effects on semantic priming in auditory word processing . Psychophysiology , 30 , 161 – 169 .

Birren J. & Fisher L . 1995 . Aging and speed of behavior: Possible conse- quences for psychological functioning . Annual Psychological Review , 26 , 329 – 353 .

Cabeza R . 2001 . Cognitive neuroscience of aging: Contributions of function- al neuroimaging . Scand J Psychol , 42 , 277 – 286 .

Cabeza R. , Daselaar S. & Dolcos F . 2004 . Task-dependent and task-specifi c age effects on brain activity during working memory, visual attention, and episodic retrieval . Cereb Cortex , 14 , 364 – 375 .

Chiarello C. , Burgess C. , Richards L. & Pollock A . 1990 . A semantic and associative priming in the cerebral hemispheres . Brain Lang , 38 , 75 – 104 .

Chiarello C. & Richards L . 1992 . Another look at categorical priming in cerebral hemispheres . Neuropsychologia , 30 , 381 – 392 .

Coltheart M . 1981 . MRC psycholinguistic database . Q J Exp Psychol , 33A , 497 – 505 .

Cool Edit Pro (Version 2.1) (computer software) . 2003 . Phoenix, USA:

Syntrillium Software Corporation .

Deacon D. , Mehta A. , Tinsley C. & Nousak J . 1995 . Variation in the latencies and amplitudes of N400 and NA as a function of semantic priming . Psychophysiology , 32 , 560 – 570 .

D ’ Esposito M. , Zarahn E. , Aguirre G. & Rypma B . 1999 . The effect of normal aging on the coupling of neural activity to the bold hemodynamic response . Neuroimage , 10 , 6 – 14 .

Gilhooly K. & Logie R . 1980 . Age-of-acquisition, imagery, concreteness, familiarity, and ambiguity measures for 1944 words . Behavior Research Methods and Instruments , 12 , 395 – 427 .

Gray S. , Plante E. , Vance R. & Henricksen M . 1999 . The diagnostic accuracy of four vocabulary tests administered to pre-school age children . Language Speech and Hearing Services in Schools , 30 , 196 – 206 . Grose-Fifer J. & Deacon D . 2004 . Priming by natural category member-

ship in the left and right cerebral hemispheres . Neuropsychologia , 42 , 1948 – 1960 .

Haug H. & Eggers R . 1991 . Morphometry of the human cortex cerebri and corpus striatum during aging . Neurobiology of Aging , 12 , 336 – 338 . Holcomb P. , Coffey S. & Neville H . 1992 Visual and auditory sentence pro-

cessing: A developmental analysis using event-related brain potentials . Dev Neuropsychol , 8 , 203 – 241 .

Holcomb P . 1993 . Semantic priming and stimulus degradation: Implica- tions for the role of N400 in language processing . Psychophysiology , 30 , 47 – 61 .

Jasper H . 1958 . Report of the committee on methods of clinical examination in electroencephalography . Appendix: The ten-twenty electrode system of the International Federation. Electroencephalography and Clinical Neurophysiology , 10 , 370 – 375 .

Juottonen K. , Revonsuo A. & Lang H . 1996 . Dissimilar age infl uences two ERP waveforms refl ecting semantic context effect . Cognitive Brain Research , 4 , 99 – 107 .

Kahlaoui K. , Vlasblom V. , Lesage F. , Senhadji N. , Benali H. et al . 2007 . Semantic processing of words in the aging brain: A near-infrared spectroscopy (NIRS) study . Brain Lang , 103 , 144 – 145 .

Kok A . 2001 . On the utility of P3 amplitude as a measure of processing capacity . Psychophysiology , 38 , 557 – 77 .

Kutas M. & Hillyard S . 1984 . Brain potentials during reading refl ect word expectancy and semantic association . Nature , 307 , 161 – 163 .

Kutas M. & Hillyard S . 1989 . An electrophysiological probe of incidental semantic association . Journal of Cognitive Neuroscience , 1 , 38 – 49 . Kutas M. & Iragui V . 1998 . The N400 in a semantic categorization

task across six decades . Electroencephalography and Clinical Neurophysiology , 108 , 456 – 471 .

Laver W. & Burke D . 1993 . Why do semantic priming effects increase in old age? A meta-analysis. Psychology & Aging , 8 , 34 – 43 .

Mehta J. , Jerger S. , Jerger J. & Martin J . 2009 . Electrophysiological correlates of word comprehension: Event-related potentials (ERP) and independent component analysis . Int J Audiol , 48 , 1 – 11 .

(8)

Markson L. & Bloom P . 1997 . Evidence against a dedicated system for word learning in children . Nature , 385 , 813 – 815 .

Molfese D.L . 1995 . Electrophysiological responses obtained during infancy and their relation to later language development: Further fi ndings . In: M.G. Tramontana & S.R. Hooper (eds.) , Advances in Child Neuropsychology , Vol. 3 . New York: Springer-Verlag , pp. 1 – 11 .

Milberg W. & Blumstein S.E . 1981 . Lexical decision and aphasia: Evidence for semantic processing . Brain Lang , 14 , 371 – 385 .

Mueller H. , King J. & Kutas M . 1997 . ERP elicited by spoken relative clauses . Cognitive Brain Research , 5 , 193 – 203 .

Picton T . 2011 . Human Auditory Evoked Potentials . San Diego: Plural Publishing .

Rabinowitz J . 1986 . Priming in episodic memory . Journal of Gerontology , 41 , 204 – 213 .

Rice M. , Buhr J. & Oeting J . 1992 . SLI children ’ s quick incidental learning of words: The effect of a pause . J Speech Hearing Res , 35 , 1040 – 1048 .

Statview (computer software) . 1998 . Cary, USA: SAS Institute, Inc . Stim2 (computer software). 2003 . El Paso, USA: Compumedics Neuroscan . Velanova K. , Lustig C. , Jacoby L. & Buckner R . 2006 . Evidence for frontally

mediated controlled processing differences in older adults . Cerebral Cortex , 31 , 2 – 14 .

Woodruff-Pak D . 1997 . The neuropsychology of aging . Malden, USA:

Blackwell .