J Am Acad Audiol 25:975–982 (2014)
Cognition and Speech-In-Noise Recognition: The Role of Proactive Interference DOI: 10.3766/jaaa.25.10.6 Rachel J. Ellis* Jerker Ro¨nnberg*
Abstract Background: Complex working memory (WM) span tasks have been shown to predict speech-in-noise (SIN) recognition. Studies of complex WM span tasks suggest that, rather than indexing a single cognitive process, performance on such tasks may be governed by separate cognitive subprocesses embedded within WM. Previous research has suggested that one such subprocess indexed by WM tasks is proactive interference (PI), which refers to difficulties memorizing current information because of interference from previously stored long-term memory representations for similar information. Purpose: The aim of the present study was to investigate phonological PI and to examine the relationship between PI (semantic and phonological) and SIN perception. Research Design: A within-subjects experimental design was used. Study Sample: An opportunity sample of 24 young listeners with normal hearing was recruited. Data Collection and Analysis: Measures of resistance to, and release from, semantic and phonological PI were calculated alongside the signal-to-noise ratio required to identify 50% of keywords correctly in a SIN recognition task. The data were analyzed using t-tests and correlations. Results: Evidence of release from and resistance to semantic interference was observed. These measures correlated significantly with SIN recognition. Limited evidence of phonological PI was observed. Conclusions: The results show that capacity to resist semantic PI can be used to predict SIN recognition scores in young listeners with normal hearing. On the basis of these findings, future research will focus on investigating whether tests of PI can be used in the treatment and/or rehabilitation of hearing loss. Key Words: Cognition, executive function, proactive interference, speech-in-noise recognition, working memory Abbreviations: PI 5 proactive interference; SIN 5 speech-in-noise; SNR 5 signal-to-noise ratio; WM 5 working memory; WMS 5 working memory span
INTRODUCTION
C
ognition refers to mental processes such as information processing, memory, and attention. Cognitive processes may be affected by interference and inhibition, which are mechanisms that disrupt the storage and/or retrieval of information in memory. Proactive interference (PI; sometimes called proactive
inhibition), on which the present study will focus, occurs when items already stored in memory interfere with the acquisition of new memories that are similar to existing representations. PI, or more precisely, the capacity to resist the effects of PI (i.e., to inhibit the influence of PI), has been the topic of many research articles across a broad range of fields. Evidence for an effect of PI has been reported
*Linnaeus Centre HEAD, Swedish Institute for Disability Research, Department of Behavioural Sciences and Learning, Linko¨ping University, Linko¨ping, Sweden Rachel J. Ellis, Linnaeus Centre HEAD, Swedish Institute for Disability Research, Department of Behavioural Sciences and Learning, Linko¨ping University, 58183, Linko¨ping, Sweden; E-mail:
[email protected] Parts of this work were presented at the Cognitive Hearing Science for Communication conference, Linko¨ping, Sweden, June 16–19, 2013. This research was funded by the Linnaeus Centre HEAD, The Swedish Research Council (grant number: 2007-8654).
975 Delivered by Ingenta to: American Academy of Audiology Members IP : 90.228.209.200 On: Mon, 22 Dec 2014 16:23:08
Journal of the American Academy of Audiology/Volume 25, Number 10, 2014
in relationship to memory for naming odors (Lawless and Engen, 1977), memory for advertisements (Unnava and Sirdeshmukh, 1994), risk for development of posttraumatic stress disorder (Verwoerd et al, 2009) and fluid intelligence (Borella et al, 2006), among other things. Traditionally, PI has been investigated using the BrownPeterson paradigm (Brown, 1958; Peterson and Peterson, 1959). The Brown-Peterson task requires the participant to memorize lists of letters or words (typically three or four) for recall after an interval lasting between 0 and 20 sec (Floden et al, 2000). The test was later modified to allow for the examination of release from PI (Wickens et al, 1963; Wickens, 1970) by manipulating the semantic categories of the to-be-remembered word lists. Release from PI refers to the benefit of performance in trials in which PI has not had a chance to build up, compared with trials designed to be high in PI. Keppel and Underwood (1962) found that the effects of PI occur both for memories stored in the short-term and long-term. The relationship between PI and working memory (WM) is somewhat more complex because WM involves the active manipulation of information, whereas the long-term and short-term memories are dominated by storage processes. As such, WM span (WMS) tests are thought to index a wide range of cognitive processes rather than simply information storage (eg, see So¨rqvist et al, 2010). According to Kane and Engle (2000), one of the primary functions of WM is to enable resistance to PI. They compared the effect of PI on performance in a memory for word lists task in two groups of participants: one that had scored highly in a WMS test (operation span) and one that had performed poorly in the same WMS test. The results showed that the high-WMS capacity group showed less evidence of an effect of PI than did the low-WMS capacity group. However, when participants were asked to complete a secondary task (finger tapping with variation in the complexity of the to-betapped sequence, and therefore, increasing the cognitive load induced by the task) at the same time as the PI memory task, both groups showed similar levels of susceptibility to PI. Kane and Engle (2000) suggested that their results provided evidence that, as increasing cognitive load did not affect performance, the low-WMS group did not normally allocate attention to resisting PI whereas the high-WMS group, for whom performance was affected by cognitive load, did allocate attention to resisting PI. The results reported by Kane and Engle (2000) provide evidence of a relationship between complex WMS and PI, suggesting that those who perform better on WMS tasks are better able to resist PI. On the basis of these findings, a number of experiments have been conducted in an attempt to further explore the nature of the relationship between PI and WM. One such experiment was reported by Friedman and Miyake (2004). The
primary purpose of their study was to explore inhibition and interference in relation to three cognitive functions. The three functions on which the study focused were prepotent response inhibition, resistance to distractor interference, and resistance to PI. They conducted a confirmatory factor analysis on data from 220 adults, the results of which suggested that proponent response inhibition and resistance to distractor interference were related; however, neither of these functions was related to resistance to PI. Subsequent structural equation modeling revealed that performance on the reading span test and frequency of unwanted intrusive thoughts were related to resistance to PI. Results by Friedman and Miyake (2004) support those reported by Whitney et al (2001), who investigated the executive functions underlying performance on the reading span test, finding that these functions fell into two broad categories, namely manipulation capacity and susceptibility to interference. These findings have been supported by results reported by Lustig et al (2001), who found that reducing the amount of PI in a WMS test was associated with increasing scores, a result also observed by May et al (1999). Furthermore, Lustig et al (2001) found that altering the amount of PI in the WMS task affected how well the WMS test score predicted performance in a proserecall task. The effect of manipulating PI within a WMS task was also investigated by Blalock and McCabe (2011), who found that PI modulated the capacity of the WMS score to predict fluid intelligence (with the high-PI trials predicting performance better than the low-PI trials). Collectively, these findings suggest that interference is an important factor in explaining the association between WMS scores and higher cognitive functions. One higher cognitive process, shown to be predicted by WMS scores, is the perception of speech or, more specifically, the perception of speech that has been distorted in some way, whether by noise, hearing loss, or hearing aid use (see Akeroyd, 2008, for a review). Given the large body of research implicating PI in WM, this begs the question of whether the ability to resist PI is related to the ability to perceive distorted speech. Studies of release from PI in verbal tasks tend to use semantic category shifts to induce such a release. Of interest is whether a phonological shift (such as that induced by hearing loss or hearing aid use) may be sufficient to induce a release from PI. As sensorineural hearing loss tends to affect the high frequencies more than the low and mid frequencies, hearing loss (and, in some cases, hearing aid use) may result in phonemic confusions due to the loss (or subsequent replacement) of this high-frequency information (eg, see Bilger and Wang, 1976; Simpson et al, 2006). As hearing loss and hearing aid processing primarily affect the phonology of speech, not the semantic content, it is possible that sensitivity to phonological PI may be a more sensitive index than semantic PI of the cognitive processes influencing
976 Delivered by Ingenta to: American Academy of Audiology Members IP : 90.228.209.200 On: Mon, 22 Dec 2014 16:23:08
Proactive Interference and Speech Recognition/Ellis and Ro¨nnberg
the degree of benefit to speech perception obtained from amplification or the amount of time required to acclimatize to a new hearing aid. To date, the effects of phonological shifts in speech material on PI have only been subjected to very limited research. One such study that sought to investigate the effect of phonological factors on PI was conducted by Coltheart and Geffen (1970). Coltheart and Geffen’s study used 36 listeners with normal hearing, who each received five stimulus triads, presented visually. Nine stimuli were used (varied in terms of manner and place /eb/ /ed/ /eg/ /ep/ /et/ /ek/ /em/ /en/ /eng/), and the participants’ task was to repeat the stimulus immediately after presentation. After the presentation of the final stimulus in the triad, a series of geometrical shapes were presented for 20 sec after which participants were asked to recall the triad in the correct order. The results showed evidence of a build-up of PI in the first four trials followed by a release from PI in the final trial as expected, and was the case for both manner and place differences. The results of a control experiment confirmed that these results were the result of the phonological nature of the stimuli as opposed to the fact that the stimuli in the final trial were always novel (confirming results reported by Wickens et al (1963), which suggested that PI is unaffected by the novelty of the stimuli). In addition, a small body of research has focused on the role of PI in pitch-comparison tasks. Ruusuvirta (2000) found that the nature of three interfering tones, presented before a pair of to-be-compared tones, affected performance on the comparison task. A later study conducted by Ruusuvirta et al (2008) also found evidence of PI in a task using successive pairs of tones, with the nature of the preceding pair affecting judgment of the later pair. Tehan and Humphreys (1995) also found some evidence of PI effects in their study of transient phonemic codes. Specifically, immediate recall of item lists categorized on the basis of whether or not they rhymed, as opposed to being categorized on the basis of semantic similarity, was significantly worse for rhyming than for nonrhyming lists. The aims of the current study were to further investigate evidence for phonological PI and to explore whether measures of semantic and/or phonological PI can be used to predict speech-in-noise (SIN) recognition scores. If such a relationship is observed, this may indicate that PI might play a role in determining how well people are able to adapt to hearing loss and, where applicable, amplification. Therefore, tests of PI could provide valuable information to help guide the expectations of new hearing aid users and in tailoring individualized support to aid rapid acclimatization. The ultimate goal would be to further the development of tests that could be used clinically to aid the treatment and rehabilitation of listeners with a hearing loss. However, in order to determine the role of PI in recognizing speech processed by a hearing aid, it is first
necessary to investigate whether PI can be used to predict speech recognition in listeners with normal hearing. METHODS Participants A total of 24 native Swedish speakers with normal hearing (14 males, mean age 5 23.8 yr old) were recruited to take part in the study. Participants were paid 500 SEK for their participation. The study was conducted in accordance with the Declaration of Helsinki. The participants reported having normal hearing, which was then confirmed using either pure-tone audiometry (a threshold of better than 20 dB HL at frequencies between 125–8000 Hz) or performance on the Hearing Bridge SIN test (see Molander et al, 2013 for details). Procedure All testing took place in the course of one session. The PI test was always administered after the SIN test. SIN Recognition Test After presentation of one practice list, six blocks of 10 sentences from the Swedish HINT corpus (Ha¨llgren et al, 2006) were presented via an external soundcard (Roland Quad-Capture Audio Interface) and supra-aural headphones (Sennheiser HD 215). Within each list, all sentences were presented at 60 dB SPL, and the level of the noise (two-talker Swedish babble) varied in 3 dB steps from 112 to –15 dB in a fixed order (in ascending order of difficulty, and thus decreasing signal-to-noise ratio [SNR]), similar to the method recommended by Wilson et al (2007). The participants’ task was to repeat each sentence back to the experimenter. The percentage of keywords correctly recognized (across all SNRs) was then calculated. In addition, as per the recommendations made by Wilson et al (2007), the Spearman-Ka¨rber equation (Finney, 1952) was used to calculate the SNR required to obtain a score of 50% correct in the SIN recognition test. The test took approximately 15 min to complete. PI Test The PI test was based on that used by Kane and Engle (2000). The test consisted of nine blocks of trials, preceded by one practice block. Each block consisted of four lists of seven words: the first three lists belonging to the same category and the final list belonging to a different category (see Appendix 1 for examples of lists used in the experiment). After the presentation of each list, participants completed a distractor task for 16 sec to prevent rehearsal. The distractor task required the participants to continue (verbally) a letter-number sequence, the first
977 Delivered by Ingenta to: American Academy of Audiology Members IP : 90.228.209.200 On: Mon, 22 Dec 2014 16:23:08
Journal of the American Academy of Audiology/Volume 25, Number 10, 2014
item of which (e.g., “E8”) was presented on the screen immediately after presentation of the final word in the list. After completion of the distractor task, the participants were given 20 sec to recall as many words as possible from the list. In the semantic condition (3 blocks), categories were based on semantic categories (e.g., capital cities or birds). In the phonological condition (6 blocks), categories were based on either initial phoneme (2/6 blocks), gender of the speaker (2/6 blocks), or the combined effect of both (2/6 blocks). Stimuli in the phonological lists were not semantically related, nor were those in the semantic lists phonologically related. All stimuli were disyllabic words presented aurally (using the same apparatus as used in the SIN recognition test) at 65 dB SPL. The order of presentation of the blocks was randomly selected as was the order of words within each list. Participants were given the option of taking a break between each block. Two outcome measures were then calculated: (1) resistance to PI (list 1 recall – list 3 recall), where a lower score indicates greater resistance to PI; and (2) release from PI (list 4 recall – list 3 recall), where a higher score indicates greater release from PI. The test took approximately 50 min to complete. RESULTS
T
he data were checked for normality before analysis. No evidence of significant skewness or kurtosis was evident; thus, paired-samples t-tests were used to test for evidence of resistance to, and release from, semantic and phonological PI. PI Semantic PI
Figure 1. Pattern of recall in the semantic PI blocks. Error bars represent 61 SE.
from semantic PI, and resistance to phonological PI. As the t-tests revealed no significant effect of release from phonological PI, this variable was not included as a predictor. The mean SIN recognition score was 67.02% correct (SD 5 8.46). The mean SNR required to get 50% correct in the SIN test was 6.52 dB (SD 5 2.53). The results showed that SIN recognition (based on the SNR required to obtain a score of 50% correct) correlated significantly with resistance to semantic PI only (r 5 0.44, p 5 0.032, see Figure 3) with associations between SIN recognition and release from semantic PI (r 5 0.14, p 5 0.514, see Figure 4) and resistance to phonological PI (r 5 –0.05, p 5 0.822, see Figure 5), respectively, failing to reach significance.
The mean number of items recalled per list is shown in Figure 1. The t-tests revealed significant effects of resistance to PI (t[70] 5 7.34, p , 0.001; effect size of r2 5 0.43) and, importantly, release from PI (t[70] 5 4.28, p , 0.001, effect size of r2 5 0.21). Phonological PI The mean number of items recalled per list is shown in Figure 2. Data were collapsed between the different types of phonological PI trials in order to increase the power to detect an effect. The t-tests revealed a significant effect of resistance to PI (t[143] 5 3.27, p 5 0.001, effect size of r2 5 0.07) but not of release from PI (t[143] 5 –1.12, p 5 0.267, effect size of r2 5 0.01). Predictors of SIN Recognition Three measures were investigated as possible predictors of SIN recognition: resistance to semantic PI, release
Figure 2. Pattern of recall in the phonological PI blocks. Error bars represent 61 SE.
978 Delivered by Ingenta to: American Academy of Audiology Members IP : 90.228.209.200 On: Mon, 22 Dec 2014 16:23:08
Proactive Interference and Speech Recognition/Ellis and Ro¨nnberg
Figure 3. Relationship between the mean SNR required to obtain 50% correct in the SIN test and mean resistance to semantic PI.
Figure 5. Relationship between the mean SNR required to obtain 50% correct in the SIN test and mean resistance to phonological PI.
The data were also reanalyzed to investigate correlations between the measures of PI and SIN performance (based on % keywords correctly identified) in the difficult SIN conditions (sentences 6–10 of each block) only. The results obtained were very similar to those calculated using the complete SIN data, with a significant correlation only being observed with resistance to semantic PI (r 5 –0.48, p 5 0.017). The associations between SIN recognition in the difficult conditions and release from semantic PI (r 5 –0.18, p 5 0.397) and resistance to phonological PI (r 5 0.06, p 5 0.795) again failed to reach significance. Thus, the data are robust and the pattern of results consistent regardless of subdivision of the conditions.
DISCUSSION
Figure 4. Relationship between the mean SNR required to obtain 50% correct in the SIN test and mean release from semantic PI.
T
he results provide evidence of semantic and, to some extent, phonological PI. Furthermore, the results show that resistance to semantic PI correlates significantly with sentence-in-noise recognition scores. Evidence for PI The results of the study provide evidence of both resistance to and release from semantic PI. Thus, the results support the effects reported initially by Wickens et al (1963) and provide evidence of the sensitivity of the test paradigm used in the present study to detect the effects of PI. The evidence for phonological PI was, however, less conclusive; the findings indicated a significant effect of resistance to PI but not of release from PI. This pattern of results seems to indicate that the participants did not categorize the presented word lists based on their phonological characteristics. That we see evidence of resistance to PI may indicate that, given the absence of semantic relationships among the word lists, participants treated all lists in the phonological PI test blocks as one category. Thus, we see a build-up of interference from one trial to the next, but no benefit to recall with a change of phonological category. This deficit in performance as the test block progresses may also be related to the phonological similarity effect (Conrad, 1964; Baddeley, 1966). If the phonological similarity effect was affecting performance, we would expect recall for lists of phonologically similar words to be poorer than for lists of phonologically dissimilar words. Given that all of the lists in the phonological PI trials consisted of words that were phonologically similar in some dimension (either in terms of initial phoneme or speaker), it is difficult to
979 Delivered by Ingenta to: American Academy of Audiology Members IP : 90.228.209.200 On: Mon, 22 Dec 2014 16:23:08
Journal of the American Academy of Audiology/Volume 25, Number 10, 2014
assess the contribution of the phonological similarity effect to task performance. However, the phonological similarity effect cannot explain why performance declined as the trial progressed. It may be argued that such decline may relate to fatigue; however, no evidence exists of an order effect on recall performance between blocks. Therefore, it seems that PI is indeed the most likely explanation for these results. These findings differ from those reported by Coltheart and Geffen (1970), who found evidence of both resistance to and release from PI in their study of monosyllabic nonsense words (categorized on the basis of manner and place of articulation). This difference may relate to the fact that nonsense words were used in their study, whereas meaningful words were used in the present study. As such, participants in the Coltheart and Geffen (1970) study had to rely purely on the phonological characteristics of the stimuli in order to recall them, there being no way to do so based on semantic information. This condition may have magnified the influence of phonological information on recall and resulted in the PI effects reported. This may also explain the difference between the results of the present study and those of studies reporting evidence of phonological PI in pitch-comparison tasks (Ruusuvirta, 2000; Ruusuvirta et al, 2008). Tehan and Humphreys (1995, 1998) and Ralph et al (2011) report the results of studies using meaningful stimuli (rhyming and nonrhyming monosyllabic words) to investigate PI based on phonological characteristics. However, they investigated immunity to PI rather than build-up and release of PI, as in our case. Immunity to PI is based on examining build-up (or the lack thereof) of PI, using a cued recall task. As such, evidence for PI in these studies is based only on resistance to PI and does not measure release from it. Thus, these results may be considered to be consistent with the findings obtained in the present study. PI and SIN Recognition The results of the study indicate that capacity to resist PI can be used to predict SIN recognition in young listeners with normal hearing. However, this finding only applies to scores obtained in the semantic PI test and not those obtained in the phonological PI test. There are a number of possible explanations for this, the most obvious being that the effect of PI was simply stronger in the semantic PI test than in the phonological PI test. It may also be the case that the relationship between scores on the semantic PI test and the SIN test may have been influenced by the type of speech materials used in the test. The sentence lists were taken from the Swedish HINT corpus (Ha¨llgren et al, 2006) and are both semantically and grammatically correct, meaning that, in some instances, semantic information could be used to fill in gaps in recognition caused by the noise. Had sentences that were grammatically correct but semantically
implausible been used, it is possible that the patterns of association with the semantic and phonological PI scores may have been different. This relationship may also have been affected by the type of noise used to mask the sentences. In this study, 2-talker semantically meaningful speech was used as a masker. Had we used a masker that did not contain any semantically meaningful information, again we may have seen different results. One may also question why SIN recognition relates only to resistance to PI and not to release from PI. One explanation for this may be that resistance to PI influences performance in immediate online recognition tasks such as that used in this study. It is plausible, however, that release from PI may relate to the capacity to learn to adapt to degraded speech, in acclimatization designs (although the effects of acclimatization are not universally agreed upon—eg, see Saunders and Cienkowski, 1997). Previous research into cognitive predictors of SIN recognition tends to have focused on WM capacity. That the results of this study provide evidence of a relationship between PI and SIN recognition may have implications for our understanding of the mechanisms by which performance on WMS tests predicts SIN recognition. The results of the present study suggest that it may be the capacity of complex WM tasks to index the effects of PI that underpin their capability to predict SIN recognition scores. This trend is consistent with the subcomponent theory of WM (So¨rqvist et al, 2010) and with previous studies that have found a direct effect of manipulating PI on complex span task scores and their capacity to predict higher cognitive processes (eg, see Kane and Engle, 2000; Lustig et al, 2001; Blalock and McCabe, 2011). Future Research The results indicate that semantic, but not phonological, PI can be used to predict SIN recognition in listeners with normal hearing. Future research should focus on attempting to replicate these findings in listeners with hearing loss. The effects of PI on acclimatization to hearing aids and/or novel signal-processing strategies should then be investigated. Such studies would allow us to determine the potential usefulness of tests of PI in a clinical audiology setting, either with respect to the selection of particular hearing aid settings or in terms of influencing the timeframe or nature of the treatment after completion of fitting of a hearing aid. CONCLUSIONS
T
he goals of this study were to investigate phonological PI and to determine the relationship between PI (both phonological and semantic) and SIN recognition in listeners with normal hearing. Evidence for phonological PI was limited and did not predict SIN recognition.
980 Delivered by Ingenta to: American Academy of Audiology Members IP : 90.228.209.200 On: Mon, 22 Dec 2014 16:23:08
Proactive Interference and Speech Recognition/Ellis and Ro¨nnberg
However, robust evidence of semantic PI was observed, with resistance to semantic PI correlating significantly with SIN recognition. As expected, better speech recognition scores are associated with a greater capacity to resist the effects of PI. These findings suggest that the capacity of WM tasks to predict SIN recognition may, in part, relate to their capacity to index an individual’s ability to manage the effects of PI. Further work will be conducted to investigate the relationship between semantic PI and SIN recognition in listeners with hearing loss. Acknowledgments. The authors thank Victoria Stenba¨ck, ¨ rjan Dahlstro¨m, and Amin Saremi for Niklas Ro¨nnberg, O technical assistance; the participants for donating their time; and the reviewers for their helpful comments.
REFERENCES Akeroyd MA. (2008) Are individual differences in speech reception related to individual differences in cognitive ability? A survey of twenty experimental studies with normal and hearing-impaired adults. Int J Audiol 47(Suppl 2):S53–S71. Baddeley AD. (1966) Short-term memory for word sequences as a function of acoustic, semantic and formal similarity. Q J Exp Psychol 18(4):362–365. Bilger RC, Wang MD. (1976) Consonant confusions in patients with sensorineural hearing loss. J Speech Hear Res 19(4):718–748. Blalock LD, McCabe DP. (2011) Proactive interference and practice effects in visuospatial working memory span task performance. Memory 19(1):83–91. Borella E, Carretti B, Mammarella IC. (2006) Do working memory and susceptibility to interference predict individual differences in fluid intelligence? Eur J Cogn Psychol 18(1):51–69. Brown JA. (1958) Some tests of the decay theory of immediate memory. Q J Exp Psychol 10(1):12–21. Coltheart M, Geffen G. (1970) Grammar and memory I. Phonological similarity and proactive interference. Cognit Psychol 1(3):215–224. Conrad R. (1964) Acoustic confusions in immediate memory. Br J Psychol 55(1):75–84.
Keppel G, Underwood BJ. (1962) Proactive inhibition in short-term retention of single items. J Verbal Learn Verbal Behav 1(3):153–161. Lawless H, Engen T. (1977) Associations to odors: interference, mnemonics, and verbal labeling. J Exp Psychol Hum Learn 3(1):52–59. Lustig C, May CP, Hasher L. (2001) Working memory span and the role of proactive interference. J Exp Psychol Gen 130(2):199–207. May CP, Hasher L, Kane MJ. (1999) The role of interference in memory span. Mem Cognit 27(5):759–767. Molander P, Nordqvist P, Oberg M, Lunner T, Lyxell B, Andersson G. (2013) Internet-based hearing screening using speech-in-noise: validation and comparisons of self-reported hearing problems, quality of life and phonological representation. BMJ Open 3(9):e003223 10.1136/bmjopen-2013-003223. Peterson LR, Peterson MJ. (1959) Short-term retention of individual verbal items. J Exp Psychol 58:193–198. Ralph A, Walters JN, Stevens A, et al. (2011) Immunity to proactive interference is not a property of the focus of attention in working memory. Mem Cognit 39(2):217–230. Ruusuvirta T. (2000) Proactive interference of a sequence of tones in a two-tone pitch comparison task. Psychon Bull Rev 7(2):327–331. Ruusuvirta T, Wikgren J, Astikainen P. (2008) Proactive interference in a two-tone pitch-comparison task without additional interfering tones. Psychol Res 72(1):74–78. Saunders GH, Cienkowski KM. (1997) Acclimatization to hearing aids. Ear Hear 18(2):129–139. Simpson A, Hersbach AA, McDermott HJ. (2006) Frequencycompression outcomes in listeners with steeply sloping audiograms. Int J Audiol 45(11):619–629. So¨rqvist P, Ljungberg JK, Ljung R. (2010) A sub-process view of working memory capacity: evidence from effects of speech on prose memory. Memory 18(3):310–326. Tehan G, Humphreys MS. (1995) Transient phonemic codes and immunity to proactive interference. Mem Cognit 23(2):181–191. Tehan G, Humphreys MS. (1998) Creating proactive interference in immediate recall: building a dog from a dart, a mop, and a fig. Mem Cognit 26(3):477–489. Unnava HR, Sirdeshmukh D. (1994) Reducing competitive ad interference. J Mark Res 31:403–411.
Finney DJ. (1952) Statistical Method in Biological Assay. London, UK: C. Griffen.
Verwoerd J, Wessel I, de Jong PJ. (2009) Individual differences in experiencing intrusive memories: the role of the ability to resist proactive interference. J Behav Ther Exp Psychiatry 40(2):189–201.
Floden D, Stuss DT, Craik FIM. (2000) Age differences in performance on two versions of the Brown-Peterson task. Aging Neuropsychol Cogn 7(4):245–259.
Whitney P, Arnett PA, Driver A, Budd D. (2001) Measuring central executive functioning: What’s in a reading span? Brain and Cognition, 45, 1–14.
Friedman NP, Miyake A. (2004) The relations among inhibition and interference control functions: a latent-variable analysis. J Exp Psychol Gen 133(1):101–135.
Wickens DD. (1970) Encoding categories of words: An empirical approach to meaning. Psychol Rev 77(1):1–15.
Ha¨llgren M, Larsby B, Arlinger S. (2006) A Swedish version of the hearing in noise test (HINT) for measurement of speech recognition. Int J Audiol 45(4):227–237. Kane MJ, Engle RW. (2000) Working-memory capacity, proactive interference, and divided attention: limits on long-term memory retrieval. J Exp Psychol Learn Mem Cogn 26(2):336–358.
Wickens DD, Born DG, Allen CK. (1963) Proactive inhibition and item similarity in short-term memory. J Verbal Learn Verbal Behav 2(5–6):440–445. Wilson RH, McArdle RA, Smith SL. (2007) An evaluation of the BKB-SIN, HINT, QuickSIN, and WIN materials on listeners with normal hearing and listeners with hearing loss. J Speech Lang Hear Res 50(4):844–856.
981 Delivered by Ingenta to: American Academy of Audiology Members IP : 90.228.209.200 On: Mon, 22 Dec 2014 16:23:08
Journal of the American Academy of Audiology/Volume 25, Number 10, 2014
Appendix 1—Examples of the Stimuli Used in Each Condition of the PI Test (Swedish-English) Bold typeface 5 male speaker Regular typeface 5 female speaker Example of Stimuli Used in a Semantic PI Test Block List 1 Skruvar – screws Kratta – rake Mejsel – chisel Fil – file Bultar – bolts Hammare – hammer Yxa – axe
List 2
List 3
List 4
No¨tter – nuts Klubba – club Borra – drill Naglar – nails Lie – scythe Sax – shears Vattenpass – spirit level
Kla¨mma – clamp Skrapa – scraper Hacka – pick Spackel – putty knife Kofot – crowbar Kniv – knife Limpistol – glue gun
Fja¨ril – butterfly Getting – wasp Skalbagge – beetle Mask – worm Spyfluga – bluebottle Mal – moth Tvestja¨rt – earwig
Example of Stimuli Used in a Phonological (Based on Initial Phoneme) PI Test Block List 1 Lite – little Lindring – relief lista – list lifta – hitch livba˚t – lifeboat ligger – is lipa – snivel
List 2
List 3
List 4
livrem – belt linje – line limma – size lindor – bandage lita – rely lingon – lingon livsstil – lifestyle
licens – license lika – similar lider – suffering linka – limp livvakt – bodyguard liko¨r – liqueur listig – crafty
ha¨ftar – adheres ha¨lsa – health ha¨ngla˚s – padlock ha¨rlig – lovely ha¨xa – witch ha¨nsyn – account ha¨gring – mirage
Example of Stimuli Used in a Phonological (Based on Gender of the Speaker) PI Test Block List 1 domstol – court facit – key Tekopp – teacup huka – crouch Ka¨ndis – celebrity gifta – married po¨sig – puffy
List 2
List 3
List 4
hetsig – heated cykel – bicycle halvo¨n – peninsula duktig – good hemort – residence bokning – booking patent – patent
ta¨vling – challenge servis – dinnerware balans – balance miljo¨ – environment diska – wash metal – metal ravin – ravine
nyhet – news do¨pa – baptize pajas – antics verksam – active centrum – center kultur – culture bomba – bomb
Example of Stimuli Used in a Phonological (Based on both Initial Phoneme and Gender of the Speaker) PI Test Block List 1 Bekant – known Besked – clearance bero¨m – praise berest – traveled Bestick – cutlery besta¨md – fixed behag – pleasant
List 2
List 3
List 4
bega¨r – request Betrodd – trusted belopp – amount beso¨k – visit bevis – evidence beslag – seizure bero – depend
begrepp – concept berlock – charm beslut – decision besatt – obsessed besta˚ – consist betong – concrete bekva¨m –comfortable
behov – needs bedrift – feat bedra – deceive beredd – ready beskydd – patronage besva¨r – trouble bege – proceed
982 Delivered by Ingenta to: American Academy of Audiology Members IP : 90.228.209.200 On: Mon, 22 Dec 2014 16:23:08
