Earworms ("stuck song syndrome"): towards a natural history of intrusive thoughts

Two studies examine the experience of “earworms”, unwanted catchy tunes that repeat. Survey data show that the experience is widespread but earworms are not generally considered problematic, although those who consider music to be important to them report earworms as longer, and harder to control, than those who consider music as less important. The tunes which produce these experiences vary considerably between individuals but are always familiar to those who experience them. A diary study confirms these findings and also indicates that, although earworm recurrence is relatively uncommon and unlikely to persist for longer than 24 hours, the length of both the earworm and the earworm experience frequently exceed standard estimates of auditory memory capacity. Active attempts to block or eliminate the earworm are less successful than passive acceptance , consistent with Wegner‟s (1994) t heory of ironic mental control.

Experimental research on unwanted cognitions covers many different types of intrusive thought ranging from the mundane (thought suppression experiments in which participants are asked to avoid thinking of a particular stimulus) to the pathological (auditory hallucinations, obsessive thoughts, post-traumatic "flashbacks").Studies have investigated how intrusive thoughts recur and are controlled (e.g., Wegner, 1994), but it is difficult to apply experimental methodologies to the appearance of involuntary cognitions and such experiences might be better investigated using observation of everyday occurrences (Reason & Mycielska, 1982), a procedure recently advocated as "cognitive ethology" (Kingstone, Smilek & Eastwood, 2008) and one that has been championed by one of the founders of modern cognitive psychology (Neisser, 1976).
Simply, an earworm is the experience of an inability to dislodge a song and prevent it from repeating itself in one"s head.Literary references to the phenomenon have been traced as far back as Mark Twain (1876).The virulent nature of the experience has resulted in catchy tunes being cited as possible candidates for memetic transmission and "infection" (Blackmore, 1999;Dawkins, 1976;Dennett, 1991).The phrase "earworm" is a direct translation of the German term ohrwurm which has recently gained widespread publicity via popular science books (e.g., Levitin, 2006;Sacks, 2007), and the internet.For example, a company producing "mobile brain trainers" as an electronic language-learning aid trades under the name (www.earwormslearning.com/), the exploratorium science and arts museum of San Francisco displays an online exhibition (http://www.exploratorium.edu/music/questions/earworm.html) and the University of Cincinatti hosts a website (http://www.business.uc.edu/earworms) devoted to the topic.
Despite the apparent prevalence of the earworm phenomenon, its long history, and the obvious popular interest in the topic, there has been a dearth of detailed scientific exploration.An introspective self-analysis of the subjective experience of musical imagery is reported by Brown (2006) and a factor analysis was carried out by Bar ss and Wammes (2009) in which involuntary musical imagery was found to be weakly related to personality factors and more strongly related to transliminality (a measure of cognitive penetrability).However, the references most commonly cited in the public media are to a series of unpublished studies by Kellaris (2001Kellaris ( , 2003Kellaris ( , 2006;; these unpublished studies are also briefly reviewed in Kellaris, 2008, pp. 847-849) who refers to a "cognitive itch".The largest empirical investigation of the phenomenon to date is by Liikkanen (2008) surveying nearly12,000 Finnish internet users.Liikkanen found that INMIs were typically reported as occurring "everyday" (33.2% of the sample) with more than 90% of respondents indicating that they experience an INMI at least every week (Liikkanen, 2008, p. 409).Music was the most commonly occurring form of involuntary cognition and the one most frequently considered to be disturbing.Ratings of involuntary musical images as disturbing were (at 15.1%) more than twice as high as those of the next most disturbing stimulus (visual images at 7.4%).This potentially links the earworm phenomenon to the symptoms of Obsessive-Compulsive Disorder in that it hints at a distressing element to the involuntary imagery (DSM-IV;APA, 2005).
Published accounts of studies on musical imagery also provide a useful source of information relevant to earworms.Experience-sampling a group of music students (Bailes, 2006(Bailes, , 2007) ) revealed high incidence of musical exposure over a seven-day period: 44% of the time students were hearing music when prompted to complete the form, 32% of the time they were imagining music and 21% both hearing and imagining music.As the participants were music students, this incidence is likely to be higher than in the general population, and the experience sampling does not explicitly target earworms but covers all forms of musical imagery.Nevertheless, this provides at least prima facie evidence for the prevalence of earworms and hence their legitimacy as the subject of psychological enquiry.It is also consistent with Liikkanen"s (2008) data, showing an association between INMIs and musicality.
The brief review above represents, so far as we are aware, the sum total of peer-reviewed data on the earworm phenomenon to date.Lack of published data on the phenomenon has not prevented opinion and speculation appearing under the guise of accepted scientific fact, however.Levitin (2006) for example, whilst accepting that "there has been relatively little scientific work done on the topic" nevertheless goes on to report that "musicians are more likely to have earworm attacks than nonmusicians, and that people with obsessive-compulsive disorder (OCD) are more likely to report being troubled by earwormsin some cases medications for OCD can minimize the effects" (Levitin, 2006, p. 155).Whilst the former claim seems to be borne out by Liikkanen"s subsequent (2008) study, the latter is not formally addressed in the scientific literature.Levitin also offers as an explanation for earworms that "the neural circuits representing the song get stuck in "playback mode"" (Levitin, 2006, p. 155) this is consistent with recent neuroimaging data on musical imagery (Kraemer, Macrae, Green, & Kelley, 2005), but does little more than restate the phenomenon in neural terms.Additionally, the brain areas that provide the irritation experienced in the presence of an earworm, and the areas active in attempting to -in Kellaris" (2001) phrase -"scratch the cognitive itch" are unknown.
Statements like these appearing in popular science books currently lack the support of published scientific data (no references are given to relevant studies).The statement regarding OCD, for example, is consistent with the rating of music as the most disturbing involuntary cognition (Liikkanen, 2008).It is, however, contrary to clinical reports of OCD patients.Documented musical obsessions amongst such patients are rare (Stein, Fineberg & Harvey, 2001;Zungu-Dirwayi, Hugo, van Heerden & Stein, 1999) although musical hallucinations may occur if OCD is comorbid with either schizophrenia or social phobia (Hermesh et al., 2005).Further statements include the assertion that the length of the song which is stuck is typically less than the capacity of echoic memory (given as between 15 to 30 seconds; Levitin, 2006; an estimate which is debatable (Crowder & Morton, 1969;Nairne, 1988;Penney, 1989)), that simple songs, commercial jingles or the theme music for television shows seem to get stuck more oftenor to affect more peoplethan complex pieces of music, that women suffer from more earworms than men, that the presence of lyrics is immaterial, that earworms usually include lyrics, that earworms are usually stereotyped and invariant in character and that they have "a certain life expectancy" (Sacks, 2007).All of these claims appear in one or more of Levitin (2006) or Sacks (2007) or are reported by either or both the University of Cincinatti or San Francisco Exploratorium websites as based on unpublished studies by Kellaris (2001Kellaris ( , 2003Kellaris ( , 2006)).
Other statements appear less didactic and more speculative.Sacks (2007), for example, suggests a continuum between the normal and the pathological (including repetition of movements or sounds in Tourette"s syndrome and the involuntary behaviours associated with frontal lobe damage as well as OCD), with "sudden, fullblown" earworms tending towards the pathological end of the spectrum.Some of these speculations are worth considering.Notwithstanding the lack of documented cases of earworms amongst OCD patients, the presumed relationship between earworms and OCD is given credence by evidence for failures of attentional inhibition in OCD (Muller & Roberts, 2005).Similarly, the suggestion that earworms are more prevalent now than in Twain"s day because of the ready availabilityand intrusivenessof music in everyday experience is consistent with musical imagery studies (Bailes, 2006(Bailes, , 2007;;Brown, 2006) It would be precipitate to run a full experimental study on the earworm phenomena since there is no clear theoretical framework in which to base such a study, and it is incumbent upon us to establish either the prevalence or importance of earworms to mental life before attempting to build such a framework (Neisser, 1976).
Accordingly, in what follows, we report an initial investigation of earworms using questionnaire and diary study self-report methods in order to provide the basic data to support a theoretical framework.Since a number of claims have been made linking earworms to OCD, the following questionnaire study both examines the prevalence of earworms and the extent to which they are experienced as intrusive or problematic.
OCD is conceptualised as a problem managing intrusive thoughts, but such thoughts only lead to OCD if they are distressing or interfering and whilst, by definition, the earworm is difficult to manage it is unclear whether the irritation occasioned by an earworm ever results in interference or distress.The survey is designed to examine the earworm in its natural habitatin everyday experienceand as such differs from studies of other unwanted zoological thoughts such as white bears (Wegner, 1994) or pink elephants (Navon, 1994) which have primarily been examined in the captivity of an experimental laboratory.It will also address some of the claims that have previously been made about earworms in the popular science literature, examining the extent to which these claims have any basis in fact.
Participants.An opportunity sample of 103 participants (64 male, 39 female), aged 15-57 (mean=26.24, s.d.=10.46) were recruited from amongst visitors to University Open Days and members of the public waiting at Reading railway station or visiting Forbury Gardens (a popular town-centre park) between 12-2pm on a weekday in August, 2006.
Materials and design.Questions (Appendix A) were intended to capture a number of dimensions to the earworm experience.Question 1 queried the interference level of earworms, the extent to which they disrupt normal functioning, and questions 2 and 9 queried emotional response.Questions 3, 4 and 11 were intended to record the form that earworms take, questions 5 and 10 assessed duration and frequency of earworms and questions 6-8 assessed the respondents" opinions of what caused the earworms and whether they found them easy to dismiss.Age, gender and musical experience were also recorded.

Results.
Demographic associations.All of the respondents recognised the "catchy tune" or earworm phenomenon.They reported that earworms lasted for a significant period, with the majority (88.2%) reporting tunes lasting over a period of hours or even longer.For formal analyses, likelihood ratios are predominantly reported here because reported importance (yes/no) is a categorical variable but other variables (e.g., duration) are ordinal.Chi-square was not appropriate because there were too many zero cells.Fifty-five participants (33 male, 22 female) reported previous musical training, but there were no significant associations between musical experience and any other variables, although there were positive relationships between the reported importance respondents ascribed to music (as recorded by participants" yes/no responses to this question) and both the reported duration of the tunes (measured by Question 5; likelihood ratio (8)=18.42,p =.018) and how troubling the tunes were considered to be (Question 2; likelihood ratio (8)=23.62,p =.003).That is, those who considered music important were more likely to report tunes lasting longer, and to report experiencing them as more troubling.A negative association between the importance of music and the ease with which earworms are controlled (measured by Question 8) just failed to reach significance, likelihood ratio (6)=12.54,p =.051.In a test for differences between groups, Mann-Whitney U test failed to find any difference in reported frequency of earworms between males and females, z =-.437, p =.662, inconsistent with Liikkanen (2008).
Parts of the song experienced.Respondents reported that if one part of the tune is experienced it is typically the chorus or refrain, (38.8% of the participants surveyed), but it is also likely to vary (45.6%).There was no significant relationship between the level of interference (measured by Question 1) and whether lyric, melody or both are experienced (measured by Question 3; likelihood ratio (8) = 11.82,p =.159).
Interference effects.For most respondents the earworms were not considered actively problematic but some respondents reported that earworms stopped them doing other things and wasted some (4.5%), or lots of time (5.5%).There was a trend towards an association between age and the interference measure, Kendall"s =-0.13, p =.087, (interference is not normally distributed so a non-parametric correlation, Kendall"s tau, is reported here) and a significant association between the interference measure and the importance respondents ascribed to music, likelihood ratio (8) = 23.63,p =.003 similar to the relationships between importance and other factors reported above.
There was a significant positive relationship between how worrying (Question 6b) and unpleasant (Question 9) an earworm was considered to be, r =.253, p <.05 but no relationship between the duration of the earworms and the extent to which they interfered with normal functioning, =.02, p =.83.There was a significant positive correlation between interference and controllability (Question 8) of earworms, however, =.28, p <.001 and an association between interference and the frequency (Question 10) of earworms, =-.232, p=.004, with participants reporting less frequent earworms also reporting greater interference from the earworms they experienced.
There were no relationships between the duration of the earworms and their frequency, or how easy they were to control, nor between the controllability and frequency of earworms (p>.05 in all cases).
Open-ended questions.The modal response to question 11 (whether catchy tunes that got "stuck" were always the same tune) was that each occasion was a different tune (47.3% of responses compared to 32.7% and 10.9% for the same few tunes and always the same tune respectively).Participants were also asked to give examples of tunes that they found difficult to dismiss.159 such songs were named, including 11 TV/film themes and 12 advertising jingles.In addition, there was one mention of "adverts generally", one of video-game tunes, two mentions of children"s music and one of Christmas songs.The remainder were exclusively pop music.Fourteen artists were mentioned by name with no song given.There was very little repetition within the list although, within artists, this sample mentioned Pink Floyd three times, Guns "n" Roses four times and Justin Timberlake five times.Ten other artists were similarly represented more than once, but no more than 2-3 times each.The activities participants employed to rid themselves of earworms were classified under one of five headings, given in Table 1.
The results of this study clarify some features of the earworm experience.
Notwithstanding Liikkanen"s (2008) report that intrusive musical imagery is the most disturbing of involuntary cognitions, we found little evidence that respondents considered their earworms actively problematic.Likewise we did not replicate Liikkanen"s (2008) report of a sex difference in the incidence of earworms, although this could simply be the result of a lack of statistical power in this survey (n = 103, 38% female) compared to Liikkanen"s internet study (n = 11, 904, 68% female) given the small effect size in Likkanen"s study ( p 2 =.0133).
Those who found the earworms most problematic were respondents who considered music particularly important, these participants also reported experiencing earworm-episodes of longer duration and harder to control than participants for whom music was of less importance.These results confirm previous speculation regarding a relationship between musicality and earworm experience, although they emphasize the subjective importance of music rather than musical skill or experience, contrary to Levitin"s (2006) suggestion that earworms are more prevalent amongst musicians per se, which confounds musical skill and experience with the importance ascribed to music.The results are also consistent with the idea that involuntary cognitions that are viewed as important are more likely to be problematic than those with neutral or personally unimportant content.To this extent the earworms seem to behave like pathological intrusive thoughts, which only become so if their content is viewed as sufficiently unpleasant or distressing.Additionally, participants who experience earworms less frequently, or consider them less controllable, also report them as more problematic.
The extent to which an earworm is considered worrying is also related to its unpleasantness rating, although there was no effect of unpleasantness on the earworm episode"s duration, implying that more worrying and unpleasant earworms differ from pathologically obsessive thoughts (contrary to suggestions by Levitin, 2006) in being relatively short-lived and, in a nonclinical population at least, do not attract obsessive thoughts are not considered particularly problematic or even intrusive (beyond the annoyance often associated with such experiences).This is consistent with Sacks" (2007) assumption that earworms have a certain (undetermined) "lifeexpectancy" which, presumably, is rarely prolonged to any great degree.Although the current data do not allow us to estimate this life-expectancy with any degree of precision, the modal reported duration of an earworm episode as over a period of hours is inconsistent with Levitin"s (2006) claim that earworms are limited by the capacity of auditory memory, however it is possible that a distinction needs to be drawn here between the period of time over which an earworm manifests itself (an earworm-episode) and the length of the tune constituting the earworm.Given that the majority of reported earworms were popular music tunes, it is unlikely that an earworm-tune could consistent of a musical piece longer than approximately three minutes.
Interestingly, the earworms reported were always familiar to the participants, no reports were given of unknown or novel tunes acting as earworms although this may be a consequence of the way in which the question was phrased ("Please list below any tunes that you can remember getting stuck in your head on previous occasions"), which may have encouraged participants to write down only "known" tunes.The incidence of TV/film themes or advertising jingles was also not particularly high, contrary to previous suggestions (Kellaris, 2008) and the tunes also seem to be virtually unique to each individual, suggesting that lists of the most potent earworms (e.g., Kellaris, 2001) are misleading.The earworms varied with regard to which part of the song was involved, but the chorus was most frequently cited.
Although the questionnaire study provides data which takes us beyond the previous purely speculative comments regarding earworms (Kellaris, 2008;Levitin, 2006;Sacks, 2007), a limitation on the current conclusions is that the reports of Study 1 are retrospective and hence, like other surveys of earworm experience (e.g., Liikkanen, 2008), subject to potential recollection bias.Additionally, the results of this study are averaged over a large number of individuals each recording their assessment of numerous earworm episodes with no possibility of examining individual earworms in-depth or determining which of the activities listed in Table 1 are effective at dismissing earworms.
To address these limitations, Study 2 examines individual earworm episodes, with participants recording their incidence, and their own response, as they occur.The advantages and shortcomings of both survey and diary data are discussed in-depth by Reason and Mycielska (1982), particularly with regard to their study of mental lapses and everyday errors.Like everyday errors, it seems most appropriate to study "everyday" earworms using both these techniques as they are to a certain extent complementary and more objective methods present difficulties given the infrequent and involuntary nature of the subject matter (see Reason & Mycielska, 1982, pp.21-23 for a discussion; Kingstone et al., 2008, and accompanying commentaries provide a more recent outline of the advantages, constraints and criticisms of this and other forms of "cognitive ethology").

Method.
Participants.Twenty-five individuals (10 male, 15 female) responded to an online advertisement posted in September 2006 and volunteered to participate.Of these, twelve (3 male, 9 female) submitted diary reports of their earworm experiences.Nine were undergraduate students, one was a postgraduate and two were non-students.One participant identified herself as a musician with over five years experience.None had participated in Study 1.
Materials & Design.Participants were given a standardized set of instructions explaining the nature of earworms and of the diary.They were sent a standardized form (Appendix B) and requested to submit (i.e., return by post) completed diaries at fortnightly intervals, that is after two and four weeks.

Results.
A total of 269 earworm episodes were recorded, averaging 1.12 episodes per week (range = 0.25-2.5).The earworms were associated with 199 different tunes.Of the 70 tunes which appeared more than once only 10 were reported as earworms by more than one individual, consistent with Study 1"s result showing little repetition of earworms either within or between individuals.The frequency of children"s songs (3), TV/radio themes ( 9) and advertising jingles ( 7) was broadly equivalent to that previously observed and, additionally, a small number of non-pop tunes were reported, including two folk tunes, one jazz instrumental and four show tunes.33% of the earworm episodes were reported as consisting of the chorus only, consistent with Study 1, 27% were reported as fragments (e.g., "end bit only", "small section", "just one line") likely to be shorter than the chorus and a surprising 28% of episodes reported involved the entire song.The average duration of an earworm episode, 27.25 minutes, was substantially shorter than suggested by Study 1.
The modal assessment of the earworm interference was that they did not interfere nor waste time (66% of the time), although the percentage of occasions in which they were judged to interfere a little but not waste time (10%), interfere a little and waste time (9%), interfere and waste some time (4%) or interfere and waste a lot of time (1%) were higher than those previously reported.Only 89 earworm episodes (33.1%) were reported as unpleasant (scores of a-d on the rating scale (86 episodes), or a score of f with a verbal statement consistent with a negative experience (3 episodes)).The majority (50.6%) were reported as pleasant (136 scores of e and 7 scores of f together with a positive verbal statement), the remainder being ungraded (31 episodes) or neutral (6 episodes).Of recurring earworms, only 25.56% were recorded as unpleasant.Figure 1 shows the number of times each earworm recurred and the recurrence rate of an earworm within seven days of its first appearance.

FIGURE ONE ABOUT HERE
Activities reported to try and remove earworms were categorized as in Table 1.In this sample, no active attempt to remove an unwanted earworm was made on 56% of occasions, earworm elaboration only occurred on 4% of occasions, specific displacement activities were attempted 21% of the time and generalized distraction techniques 12%.Figure 2 shows the effectiveness of three of these strategies (nothing, specific displacement and generalized distraction).Earworm elaboration is not included because the small numbers of episodes involved render this data-point unreliable although, numerically, this condition resulted in the longest earworm activity (47 minutes).

FIGURE TWO ABOUT HERE
A by-participant analysis of variance (ANOVA) was not possible because of the small numbers involved, the extent of intra-participant variability and its interaction with inter-participant variability.For example, not all participants varied in their report of distraction activities intended to displace the earworm and some participants reported multiple different types of distraction activity (hence distraction activity would qualify as both a within-and between-participant factor in such an analysis).However, a by-item analysis (Fisher, 1935) with distraction activity and participant as fixed factors and each individual earworm episode constituting an observation, found no main effect of activity, F<1, but a significant main effect of participant F(11,192) = 10.03,p <.001, p 2 =.365 and an activity by participant interaction, F(12,192) = 2.02, p =.024, p 2 =.112.The absence of a main effect is explained by missing data-points for some participants, not all of whom reported attempting distracter activities.Since only three means are involved, it is possible to run post hoc analyses on this interaction without affecting the family-wise error rate (see Howell, 2007).Post hoc Fisher"s LSD showed significant differences between no activity and both specific and generalized distraction, p =.032 and p =.048, respectively, but no differences between specific and generalized distraction activities, p =.884. Discussion.
The diary study largely confirms the findings of the questionnaire concerning both the frequency and individuality of earworms.Earworms are both frequent (although slightly less so in the diary study than in the questionnaire) and idiosyncratic, with little overlap between individuals and little recurrence (at least over the short-term) within individuals.As before, although popular tunes predominated, there was no evidence that simple songs such as children"s songs, jingles or theme tunes were over-represented in the earworms reported.The length of the earworm is frequently longer than the relatively short duration of auditory memory, with whole songs sometimes being experienced.This fact can be explained by the finding that participants in this study could name 99% of the earworms they reported.Hence, consistent with Study 1, the earworm is a familiar tune and presumably involves not just auditory memory but the reactivation of long-term memory representations.
The diary study also provides the first opportunity to consider cognitive control strategies related to earworms.The most obvious theoretical framework to apply is Wegner"s (1989Wegner"s ( , 1994) ) theory of ironic mental control.This theory suggests that the need to consciously monitor a mental control process to determine its success is self-defeating when the goal of the control process is to remove an item from conscious awareness as the representation of this goal necessarily reinstates the item in consciousness whenever the monitoring process is deployed.Hence, according to this theory, any conscious attempt to displace or remove an earworm is unlikely to succeed as monitoring this attempt simply re-presents the earworm to the individual.
The results of the by-item analysis are consistent with this idea, showing longer earworm-episodes when specific attempts are made to displace them.However, these results are equally consistent with the possibility that only particularly virulent (in some sense) earworms attracted this kind of activity and future experimental studies, with more control over the situation and running by-participant analyses, are necessary to confirm these results and decide between these two possible interpretations.
General Discussion.
With regard to the comments made about earworms in popular science sources, a number of claims were identified earlier.These were that musicians are more likely to suffer from earworms than non-musicians, that OCD patients were more likely to report earworms (and that medication for OCD can result in the disappearance of such earworms), that the length of the song which is stuck is typically less than the capacity of echoic memory (given as between 15 to 30 seconds; Levitin, 2006), that simple songs, commercial jingles or the theme music for television shows seem to get stuck more oftenor to affect more peoplethan complex pieces of music (Kellaris, 2001(Kellaris, , 2003)), that women suffer from more earworms than men, that the presence of lyrics is immaterial, that earworms usually include lyrics, that earworms are usually stereotyped and invariant in character and that they have a certain "life expectancy" (Sacks, 2007).Whilst we are not in a position to decisively refute or confirm all of these claims, nevertheless the current study has provided sufficient data to address some of these suggestions and, in some cases, to provide potential alternative interpretations of some of these claims and to point the way for future research.
Firstly, whether musicians are more likely to suffer from earworms than nonmusicians.Study 1 found no evidence for this, although this study sampled visitors to the Department of Psychology and other members of the general public, it did not specifically target professional musicians so it is unclear what results would be obtained if a specific comparison was carried out between experienced musicians and those with no musical training.The study did find positive relationships between the importance participants ascribed to music and a number of different dimensions of the earworm experience, however (including the reported duration of the earworm episode, how troubling it was considered to be, and how much the earworm interfered with other activities).These results suggest that receptiveness to music, rather than musical skill or experience, might be the appropriate dimension along which to look for musician-nonmusician differences.The current study also failed to find any sex difference, in contrast to Likkanen ( 2008), which we have interpreted as the result of a lack of statistical power to reveal a small effect size.Since gender could plausibly interact with emotional reaction to music, any future demographic study needs to take both these factors into account.
Secondly, the supposed link between earworms and OCD.We do not have formal data enabling us to examine the incidence of earworms amongst OCD patients alongside that of comparable individuals in the general population so we have no definitive confirmation or refutation of this link.Nevertheless, the data we have are suggestive.Unexpectedly, many earworm episodes were not considered unpleasant at the time they were experienced (Study 2).Thus the content of the earworm is very different from that of a pathologically obsessive thought which, by definition, is distressing.Earworms ordinarily induce far less anxiety (Study 1).It is also noticeable that, unlike an obsessive thought, a particular earworm episode was unlikely to appear more than once in a single day and was usually gone completely by the next day (Study 2).When steps were taken to try and remove the earworm these proved counterproductive in terms of the episode duration.Thus, although ironic processes of mental control (Wegner, 1994) may be invoked when attempting to rid oneself of an earworm, the varied nature of the emotional reaction, the familiarity of the earworm and its lack of recurrence over time all suggest that the earworm as a phenomenon may be linked to obsessive thought disorders less by common origins in the generation of the intrusive thought than by use of common strategies to control the episode.We suggest, therefore, that (unnamed) medications for OCD (the majority of which are antidepressants which impact upon the emotional reaction to obsessive thought), are unlikely to have the beneficial effect upon the appearance and recurrence of earworms claimed by Levitin (2006).Two reported cases in which a patient presented with distressing musical pseudo-hallucinations (Praharaj et al., 2009) or obsessions (Zungu-Dirwayi et al., 1999) proved resistant to psychopharmacological treatment.This remains an avenue for future research.
Thirdly, the length and nature of the earworm.It is unclear what the 15-30 second echoic memory limit provided by Levitin ( 2006) is based upon.The limit of the classical precategorical acoustic store (Crowder & Morton, 1969) is considerably less than this (decaying after 2-3 seconds, and with a capacity of only one item).The longer time-frame may refer to the acoustic code which Penney (1989) suggested aided memory in free recall and which she hypothesized as capable of retaining five items and lasting, in the absence of attention, for 20-60 seconds.A further possibility is the supposed 1.5-2s time limit of the "phonological store" although this is generally taken to refer to speech sounds only, and requires active rehearsal to hold in memory (although see Baddeley, 2000, for evidence on the insufficiency of this concept).All four estimates are considerably below the reported duration of an earworm episode as lasting approximately 27 minutes (study 2) or several hours (study 1), although a possible distinction between the length of an earworm-episode and an earworm-tune might account for this discrepancy.Such a distinction does not account for the report, by some participants, of whole songs (of presumably several minutes duration) appearing as earworms but the observation that in both studies only familiar songs are reported as earworms complicates the issue by implicating long-term memory in the appearance of earworms.Participants in Study 1 may have been responding to the wording of the question (which implicitly assumed that the earworm was already known) when reporting only familiar tunes as earworms but the finding was replicated in Study 2 when participants were explicitly asked "Can you name the earworm?" The tunes which participants reported as earworms also varied considerably.
Since all were known to the participants, it seems that familiarity with a tune is prerequisite for it to form an earworm (a related observation has also been made regarding musical hallucinations; Warner & Aziz, 2005), prompting the suggestion that only overlearned tunes are available to be "replayed" as earworms.Simple and repetitive tunes are more likely to become overlearned over a short time period, consistent with Kellaris (2001), but exposure to the tune is the critical factor rather than the characteristics of the tune per se, suggesting that differential musical exposure would result in different patterns of earworms.This idea is supported by the observation that although popular, and predominantly recent, music tunes (likely to be overlearned) were by far the most frequent earworms reported in the current study there was very little evidence for overlap in earworms across participants (Study 1) or even within the same participant over time (Study 2).These results do not support the idea that certain tunes are intrinsically more likely to become earworms than others but do suggest that popular or otherwise very familiar tunes interact with some other, unknown but idiosyncratic factor, to form earworms.Bars are standard error.

Figure 1 :
Figure 1: Frequency of earworm recurrence.The upper panel shows the frequency

Figure 2 .
Figure 2. Mean duration of an earworm episode as a function of distracter activity.
Figure Two