an blog | AJBlog Central | Contact me | Advertise | Follow me:

Is perfect pitch a link to autism?

We have been sent a piece of statistical analysis of a very small sample group – 16 professional musicians who can identify any note by ear, 18 who cannot and 16 non-musicians.

The study, funded by Danish institutions, finds a higher level of autism among the first segment.

Whether this is in any way significant has yet to be determined, but you might like to read a short summary here.

Related Posts Plugin for WordPress, Blogger...


  1. Robert Fitzpatrick says:

    I believe that the answer to the caption is no, based on the summary attached. But, I do believe that there is a link with certain personality traits as the synopsis indicates. I worked for many years in a place where c.50% of the inhabitants had AP (students and faculty) and many of those same people had a so-called “photographic memory.” One thing that most had in common was the serious study of either piano or violin from a very early age (between 3-5 years old).

    I know of one very famous pianist who was the son of a well-known violin teacher. The son started on the violin around age 4, but after a year, the father told him that his ear wasn’t good enough for the violin and he switched to piano and the rest, as they say, is history. This person has AP to a high degree and is extraordinarily intelligent but shows no signs of autism.

    Very interesting topic. I’m sure others will pitch in, absolutely.

    • I’ll pitch in (haha)

      We ought to take note that they didn’t control for music being a factor, i.e there wasn’t a group that had AP, but weren’t musicians. I assume the 16 non-musicians did not have AP (or were not tested for it, so some might have had it but didn’t know…). So, is there a correlation of autism : (music + AP) or can we eliminate music and just identify the correlation as autism : AP? (causation of course cannot be concluded)

      AP might help someone (autistic or not) develop their musical ability quicker, that’s all.

  2. The autism spectrum is wide. The people in my family who are on some part of the spectrum (my three brothers) all have absolute pitch. There are also people in my family who are not on the spectrum who have absolute pitch. I know non musicians who are on the spectrum who have absolute pitch (and don’t even like music), and I know plenty of musicians who have excellent instrument-specific pitch memory. Knowing a pitch on the violin, for example, has a lot to do with factors aside from how many cycles per second something is vibrating. Knowing a pitch on the piano by its resonance is not uncommon. A really good ear can certainly be developed, but I firmly believe that absolute pitch memory is hard wired and inborn. I have tried all my life to acquire it but all I can depend on is my relative pitch. I seem to have missed out on whatever part of the genetic material that causes perfect pitch and the material that causes autism.

  3. No such thing as ‘perfect’ pitch, just ‘remembered pitch’. Whole premise of this study is therefore about memory not music.

  4. Interesting point you got Elaine about the difference between a perfect pitch and an instrument-specific pitch memory. I have never thought about it that way and I´m not shure witch group I belong to. If someone asks me about notes played by another instrument, I’ll immediately translate it into fingering on the piano before giving the answer.
    About the high concentration of pianists or violinist with AP; when I was studying at the conservatory, I remember that 4 out of 6 of the pianist that year had AP, but it´s worth mentioning that the ear-trainning in musical institutions are almost always done by a techer playing on the piano, I wonder if the results would be much different if the students would have to listen to a guitar instead.

    • Robert Fitzpatrick says:

      There is no question that we are talking about a highly developed aural memory whether pitch or timbre related. As a former clarinetist, I can almost always recognize the note played on any clarinet (Bb, A, Eb, whatever) by the specific color or timbre of that sound based on my experience with the instrument. But, I can not, with absolute certainty, decipher the pitch (especially if the recording is not at A=440 which is my personal point of reference). I know of gifted children who learned on a piano tuned a half-step low who remember that reference and then have to adjust in later life.

      The question about the guitar is relevant because many of our conservatory students have a pitch memory related specifically to their own instrument. Usually, if they sing the pitch, they can identify it, but most of their aural clues come from the sound of their own instrument and the relationship between sight and touch to the aural experience. Many pianist sing as they play and thereby make that vocal-aural connection. I know a few who sing so loudly during performances that the audience thinks that some fool among the public is singing along. More than once I have been asked by an irate audience member to find the idiot singing along with the pianist and to make them stop!

      To the contributor above who says that there is “no such thing as ‘perfect’ pitch” I reply that there are those who remember pitch perfectly to the vibration or even to the ‘cent’ which is one hundreth of a semi-tone. Those rare individuals have absolute perfect pitch. Highly developed relative pitch (remembering the A or some other reference tone, and calculating all other tones as intervals based on that, for example) is actually rather common among instrumentalists at the highest level.

      I find the comments by Elaine Fine especially interesting. Autism and its other related syndromes such as Asperger’s are statistically more prevalent in males, while absolute or perfect pitch memory seems about equally divided between the sexes (or maybe that’s because many women have greater opportunity to excel in music now). If I were a scientist, I would want to study her family closely to try to discover the genetic differences, if any, among the siblings.

      Final fly in the oinment: do those with absolutely perfect pitch memory have a well-tempered ear or do they hear with “just” intonation. Pianos are tuned to a tempered scale, most great string players and singers adjust the intervals according to the needs of the tonality in which they are performing. I know of one great artist who, upon hearing questionnable intonation, say the performer is employing “expressive” intonation.

      When punctuating additional entries, I shall employ the comma of Pythagoras as I listen to the Music of the Spheres.

  5. Stephen Carpenter says:

    This is very interesting. The brain is the key it seems and the more we discover about it and how it works, the more we find we don’t know. Interesting link to photographic memory.
    Just a couple of days ago there was posted the interview with Ms. Montrero. Improvisation is another brain activity. I watched the interview and the Brahms clip a few times. Amazing to me was that her improvisations seem to have an over-arcing form to them.
    We can say what we will but how the brain functions and people do what they do has to be a source of constant amazement and enrichment.

  6. People who are native speakers of tonal languages (like those in the Chinese language group) have an absolute pitch competency, so there are strong indications that it is an acquirable (with deep immersion) rather than innate skill. Perhaps an enhanced capacity for Autism-spectrum persons to acquire absolute pitch is due to a general capacity for deeper immersion in sensory input than to a specialized capacity for absolute pitch.

    The American pianist David Burge has offered a course in absolute pitch, so there are some indications that adults can acquire the skill as well with focused training.

    • Perhaps the pitch spectrum is as varied as the autism spectrum. I’m pretty sure that pitch recognition has a lot to do with other senses. When I took ear training classes that used the piano (tempered tuning), I always had to sit in the same place in the room in order to feel I was making any kind of progress. When I hold an instrument in my hands (and I play a lot of instruments), I can find pitches easily. I can even sing a pitch I am about to play and be correct, even if I am holding a baroque flute, pitched at A=415 in my hands. I can often hear orchestral music in my head that ends up being in the proper key, but I would be hard-pressed to tell you what that key might be.

      I know people who have perfect pitch (they could tell me what key that orchestral piece was, and where in the series of key relationships it had it happened to be, or they could tell me what pitch happened to be wrong in a cluster chord I might play on the piano) who are so confident about their pitch memory that they don’t really listen to their intonation when playing string instruments, and therefore play out of tune.

    • Joanna Wesling says:

      I read a very interesting article in the Scientific American Mind back in 2010 about Language and Music, the article is called Speaking in Tones. There was a very interesting bit about perfect pitch related to tone languages- quick quote ‘Perfect pitch is remarkably common in speakers of tonal languages. Ninety-two percent of Mandarin speakers who began music lessons at or before age five had perfect pitch, compared to just 8 percent of English speakers with comparable music training.’ The study does seem to show that starting music before the age of 5 was a large influence on developing perfect pitch. Fascinating article, unfortunately not available to link, you have to pay for it!

  7. Pitch cannot be ‘perfect’ because, for practical musical purposes, it has never been constant. Of course, until the electronic ‘fixing’ of A=440 Hz no two ‘A’s were the same; they might even be different from adjacent town to adjacent town, never mind city/region/country. Bach’s ‘A’ was at least a semitone different to ours. Even now, no two orchestras are ever tuned to the same pitch (except perhaps coincidentally) because the oboe that sets the ‘A’ will not be precisely fixed to A=440, nor will each individual player hear that A at precisely the same frequency. Even the temperature of the room can adjust the frequency of A. Every single note played by a violin wanders several cents away from the ‘agreed’ orchestral pitch and there may be two dozen violinists in a string section all doing likewise: what we hear in an orchestra is an amalgamated collection of pitches which coalesce around a particular focus of pitch frequency, which may change during the course of a piece, let alone concert. Nor, incidentally, is a rigidly fixed pitch in fact desirable. The first generation of synthesizers were the first mass-market instruments to reproduce totally fixed-pitch waveforms and very soon after their emergence manufacturers and designers were seeking ways of modifying, distorting and detuning those waveforms to make them warmer, more interesting, more harmonically rich, because the pure, fixed waveforms were quickly found to be dull on the ear. Robert Fitzpatrick rightly asks what tuning system one has adopted anyway: I would guess fewer than 1,000 people on earth, whose jobs are specifically to do with the performance of early music and the making of replica instruments, can aurally identify in a blindfold test Werckmeister from Valotti, Kirnberger I from Kirnberger III etc and in what region of Europe at what time such tunings might have applied. (Admittedly, a high proportion of that 1,000 may follow your blog, Norman!) Play an instrument built before eg 1850 and it doesn’t stay ‘in tune’ with itself for longer than 10 minutes anyway, such is the sensitivity and fragility of the materials under stress. It is misleading to talk about ‘perfect’ or ‘absolute’ pitch, merely a ‘reference’ or ‘remembered’ pitch. That people on the autistic spectrum – or not – can remember pitches accurately is without doubt. But what we are discussing here is the memory function of the brain. We may as well be analysing how many people can accurately reproduce the pitch of a particular train whistle or factory siren.

    • Thank you for the precision, Howard. 998 of the 1,000 are followers. The other two are composing.

    • Hi Howard,

      Not that I’d want to pedantically correct someone as awesome as you (if you are *the* HG) but:
      “The first generation of synthesizers were the first mass-market instruments to reproduce totally fixed-pitch waveforms” – not precisely true, as a waveform doesn’t give pitch (frequency does), and in any case the early synths were notoriously bad at keeping in tune (and pleasingly bad at keeping to a perfect waveform). Only with the much later advent of digital synths did this get solved (sort of – I have an early model Yamaha DX7 which I swear does not stay in tune with itself or anything else)

      I think it’s clear that “perfect pitch” (i.e. the ability to perfectly remember a pitch) is something that being on the autistic spectrum simply makes easier, rather than being an autistic trait itself. I also suspect that being a musician or a speaker of a tonal language also makes it easier. And so being an autistic musician must make it easier still. And imagine being a Chinese autistic musician!!

      My viola teacher (a retired pro orchestral player) has excellent remembered / perfect pitch. Whenever he criticises my pitching ability (which is not infrequently!), I usually claim to be playing in Werckmeister tuning :-)

      • Hi Tim, yes, I should have phrased that “The first generation of synthesizers were the first mass-market instruments to attempt to reproduce totally fixed-pitch-frequency waveforms” . I remember playing a (then) top of the range Oberheim OBX-a on a cruise ship on the Thames in the early 80s (accompanying a not-surprisingly alarmed Rowan Atkinson) and the variation in the voltage provided by the ship’s generator had the effect of changing the key of the voltage-controlled oscillators in the OBX-a about every 30 seconds. You are right they didn’t stay in tune, but in theory they COULD, which no other instrument – especially their parent the organ – achieved, nor attempted to achieve.

  8. This is a wonderful conversation. I, however, come at this topic as a layman. I have heard anecdotes of musicians who cringe in some level of physical pain when hearing a note out of tune. If “perfect pitch” is “remembered pitch” – and it does seem logical – how can this physical response be explained? Are we speaking simply of musicians more sensitive to sound?

  9. Michael Haslam says:

    Not only was pitch as variable in Bach’s time as Howard has already described but there were different pitches between the court and the attached church, so on occasions when the orchestra played in church with the organ the organist would have to transpose down a tone (or play from a transposed part). Without the benefit of aircon and central heating the pitch of eg the church organ would differ as much as a semitone between winter and summer. Add in the fact that Bach transposed several preludes and fugues to fit his scheme for the “48″ and his Magnificat exists in two versions, one in Eb and one in D, and you have a place where for Bach and his contemporaries “perfect pitch” was meaningless.

    • . . . and flutists traveled with an array of middle joints, and used head-joint screws, and extendable foot joints to compensate for these differences.

      • It's That Steve Again (ITSA) says:

        …Hmmm. So Dr Frankenstein could have saved himself a lot of trouble by going down to the local music shop, and bolting together a flutist…

  10. Sound recognition is a sensitivity that is developed. Musicians as well as people who do not play (and do not even listen to music) can identify (i.e. remember) specific voices, and not just the voice print, but the voice in various situations of stress and joy, and many of us can tell when someone we know is trying to sound like somebody else. We are thrilled when we hear people do this well.

    Concerning tonal languages, they are spoken and not sung. Because we all speak, and because the human voice functions between a few physiologically-limited boundaries, it is easy to develop a physical connection to the spoken sounds we hear. Plus, in conversation with another person who is using five tones, it is easy to stay on track, pitch wise (though remembering which “guy” is the chicken and which is the street takes skill). When the voice is used as a musical instrument, it is used differently. We can kind of relate physically to the sounds that are being made by a coloratura, the way we can kind of relate to Olympic gymnasts or ballet dancers: most of us cannot reproduce either activity outside of our imaginations. That’s one reason we like to listen and watch.

    A note is not just a note. It has many colors that can be changed by the air, the bow arm, the fingers, the vibrato, and the tongue. There are even slight variations with open strings on a bowed instrument, which anyone with a tuning device can observe. That’s because pitch is a living thing. The fact that someone can tell a harmonica from a train whistle is a miracle, really.

    Then, every note functions in an exponential number of ways in context. The A the oboe sounds, for example, when an orchestra tunes functions is totally different from the A that the violins hold at the beginning of Mahler’s first symphony, aside from the fact that it is in a different octave. An A placed on top of a B flat major triad is far different from the A that one might play after a G sharp at the end of a phrase in the key of A minor. Placing an A atop an G sharp minor triad creates some serious dissonance, which, in fortissimo played by group of brass instruments might spell danger. Played by a group of muted strings, and resolved promptly, it could evoke some sort of pathos, and played with space between the octaves by a group of winds, with the A in the oboe, it could also be humorous.

    I believe that the beauty of music is in its relativity, contextual and otherwise. People with absolute pitch memory need to develop their relative pitch and their sensitivity to all in music that is not absolute in order to become better musicians.

  11. It's That Steve Again (ITSA) says:

    Indeed a fascinating series of conversations. I’ll not say much because I’m out of my depth when it comes to music, but on things like studies of genetic inheritance of behavioural traits, by far the biggest difficulty appears to be accounting for confounds, especially socialisation and enculturation, themselves issues which require unpacking.

    There is always a history to any human phenomena, and by the time we pay attention to such phenomena, much of that history has passed by undocumented, or poorly documented. I.e. the variables we are now interested in (which would include reinforcing contingencies) were not even considered during any historical, or diarising, documentation, so were simply not captured for consideration.

    Most such studies are small, as in the example used by Norman to stimulate this discussion. And almost all are retrospective and cross-sectional (large, prospective studies which are a combination of cross-sectional and longitudinal being difficult to achieve for various reasons). All of these impose major methodological limitations, especially on accounting for confounds, and limitations on the degree of extrapolation beyond the sample. And the sampling itself tends to have built-in biases. D.T. Campbell’s 1966 book “Experimental and Quasi-Experimental Designs for Research” still ranks among the books to study on such issues.

    Memory of course, while seeming veridical to us, is in fact reconstructed, as demonstrated well by people like Elizabeth Loftus, and to some extent the writings of Daniel C. Dennett, who has tried to pull together such threads in explaining consciousness, and whose own works keep stimulating new research by others.

    • Robert Fitzpatrick says:

      OMG, IMHO, ITSA is speaking in tongues. Please translate for old retired musicians like moi. Very impressive response, sir.

      • It's That Steve Again (ITSA) says:

        Cheers Robert. I’ll try to do justice to your query later on. I’m about to run out of time, so all I could do was try to write something to cue me in later.

      • It's That Steve Again (ITSA) says:

        Right. Sorry about the delay. I set myself a challenge I may not be up to, but we’ll see.

        I found and perused a couple of papers on this issue, but whether I get to it in this part of the dialogue I don’t know. I realised that if I wait until I’ve done enough to do justice to Robert’s query, I’ll probably never get around to it. So let’s just go on a journey and see where it takes us.

        I don’t necessarily know what the hell I’m on about, but for brevity, I will try to write as if I do. I acknowledge of course that I’m out of my depth regarding music. On behavioural and cognitive science matters etc, I’m not out of my depth, although I may be rusty, although I have been relentless in my continued pursuit of understanding – but not at the level of explaining that understanding, so this is a challenge. But how I tackle such challenges is to simply go on a journey.

        So. We have music perception, and we have the sensory apparatus facilitating that perception. Even that sentence of course unpacks into whole disciplines, let alone volumes of writing. But it’s probably an unproblematic sentence that for our purposes we don’t need to bog ourselves down unpacking, because whether I’ve worded it well or not, all the readers here will know what it attempts to portray.

        We have the ever present nature vs nurture question, which we’re all familiar with in some form (but which I note has been itself brought into question, but hey, that’s just a self-prompt for me to track down a book which I saw years ago, the title of which completely eludes me, but the subject of which doesn’t). We know that in essence the questions deriving from this pertain to how much is “innate”, and how much is learned.

        Attempting to tease apart these factors, and ascertain the degree to which they overlap and indeed influence each other is where it gets messy, but where people get to build whole careers delving into. Here the issue of confounds looms very large, and indeed is all but unavoidable. This pertains to identifying cause and effect, and being able to conclusively say that all the variables have been accounted for. It can be very difficult to devise experiments or studies of any kind which can conclusively and uncontroversially claim to have accounted for any and all confounds.

        The fields where we find the best controlling of confounds are typically say, chemistry and physics. But even there, we find controversy. Particle physics for example – which is held as a benchmark of scientific rigour, relies on statistical analyses of data, which might be graphic presentations of dots and streaks from particle collisions, and separation of “signal” from “noise”. And at that extremely fine level of measurement, you also have to eliminate confounds created by all the machinery involved, whether particle generators or measuring equipment. Andrew Pickering wrote some interesting sociological commentary on this stuff, in various articles, and a book called “The Hunting of the Quark”. On writing this, I think chemistry probably fares better. But you get the point. There are fields where variables can be and are much more accounted for than we can get in the study of the human condition.

        If we’re looking at the question of nature vs nurture, we have to try to separate learning effects. Straight away it gets messy, as learning effects begin as soon as the organism starts taking in data from the environment. And reinforcement contingencies begin immediately. If we have a child at say, three, or four years of age, or younger, whose behaviour captures our attention – in this case music propensities – this didn’t just occur in a vacuum. A whole smorgasbord of environmental input has occurred right from birth: and most of this has probably not been documented in methodologically useful sense. The older the child, the more environmental input, and the greater the number of potential confounds.

        The brain is of course growing rapidly, probably easiest portrayed by J.M. Tanner’s widely published growth curves, e.g:

        This brain growth of course comprises significant generation of neurons (neurogenesis). But it also entails significant neuronal death. Donald O Hebb described the process well, and essentially it’s natural selection at the neuronal level. Those neuronal networks receiving stimuli tend to remain, and those not receiving stimuli die off. This is why wide exposure to a range of environmental stimuli is so important for developing children regardless of their propensities.

        So brain wiring begins immediately, with significant development occurring in the first few years. And this wiring is inextricably related to environmental stimuli and reinforcement contingencies.

        What a doting parent pays attention to, and what a behavioural or cognitive scientist would want documented, are not necessarily the same thing. But by the time these folks get to study any given child, they have to make do with whatever – typically highly fallible – information is available about the child’s background to that point.

        So straight away we have confounds. Even twin studies, which are typically the best benchmarks we can find, have confounds, unspoken – and not always sound – assumptions, and the problem of small sample size. I took apart a couple of studies years ago, not for any reason other than trying to get a handle on the subject. What I found was not clarification, but unspoken and questionable assumptions, small sample sizes, and a disappointing lack of rigour from one widely acclaimed researcher for starters. What I concluded was that I was seeing ideology masquerading as scientific methodology (this is what we want to be true, and this is what our studies will show), and I left it at that.

        On sampling, it’s not easy to come across studies that comprise truly random samples of sufficient size that you can extrapolate to the general population: partly due to cost limitations, and partly due to simple ease of access. A great many subjects in studies are university undergraduates. A great many studies arise from someone’s need to complete a PhD thesis: and PhD theses are themselves just a vehicle to completion of a doctoral degree, with the primary intention of demonstrating to those marking the thesis that one can DO research: it’s seldom about attempting to comprehensively review a field and everything related to it (exhaustive reviews being not just rare, but probably non-existent) then devise a study building on this: this is not to say there’s not some good study and very good reviews. But you have to review the reviews to get a handle on what’s captured and what’s left out. I should note that August Comte complained about the difficulty of obtaining a comprehensive understanding of anything as far back as the 1800s: and information has grown exponentially since then.

        Other issues with sampling can include things like who participated and who dropped out, and whether such dropouts affect the generalisability of the study results – sometimes the missing data is where the information is. It’s just messy.

        If we are relying on the memory of people to fill in the background of a child prior to the point of study, we have the problem of what it was that people think they paid attention to, what they actually paid attention to, and the reliability of their reconstruction of events. Whereas we tend to think of our memories as accurate depictions which we can reliably retrieve from storage as a veridical copy of reality, in fact not only are our memories reconstructed, but our perception of reality itself is a construction: a perceptual trick of the brain. A very good trick that works in practice, and which correlates well enough to reality that for all practical purposes we can consider it veridical. But it’s not, and we get to see this when the brain malfunctions, or breaks down, etc.

        Oliver Sacks has written well on matters illuminating this, in such readable books as “The Man Who Mistook His Wife For A Hat”.

        Daniel C Dennett is a contemporary philosopher with a practical bent (advising on projects like Rodney Brooks’ Cog project, with related projects from his lab like Kismet, a fascinating project addressing some learning variables: , who has attempted to elucidate the problem of consciousness, including a seminal book called “Consciousness Explained”. It’s a bit of a read, and one I’ve yet to finish, despite starting on it years ago – I have a few books like that (I tack from topic to topic, and back again, then articles on the books, etc, as I work on whatever understanding of the world I am going to achieve prior to death). So I’m not going to pretend that I have a handle on it, let alone a handle to the extent that Dennett has.

        But in essence, he reckons consciousness is an evolutionary trick, but a trick that works (Frankly, I think this is quite neat: here I am blundering about the world with what looks like veridical perception, and a fairly good memory, and it’s all just a trick of a brain that could break down at some point in time and just turn to shit – I better keep at it while the going’s good!). The question then becomes “how the hell DOES it work?”, and that’s what keeps lots of people in interesting jobs.

        Music perception is central to some of these studies as we know. And also to philosophical inquiries. Has anyone here read Douglas Hofstadter’s book “Godel, Escher, Bach: An Eternal Golden Braid”? If so, what do you make of it? He writes for example of the perception of infinitely rising pitch. Some of you folks will be familiar with this regardless of whether you’ve even heard of Hofstadter.

        Checking YouTube clips of Dennett (which is a first for me), I see one where he refers to Brian Felson’s symphony “View From The Stranger’s Gallery”.

        I know nothing of this yet. But I’ll bet some of you folks do.

        Now that I’ve discovered Dennett on YouTube, I’m going to cheat and check a few out (as opposed to just wading through his books, which I’ll continue to do anyway). But it’s early days, so I can’t say which links are the best representations. I can point to this one for starters:

  12. This wonderful discussion clearly portrays the need for this topic to push forward into more research. ITSA’s remark is on point. I am a music therapist, educator, and disability rights (Neurodiversity) ally. I am looking forward to joining efforts with others in future research. Our “pool” is growing, and larger studies are imminent. This could be a matter of organizing music therapists (who are doing research) to their local autism communities, which is part of the work I do with my nonprofit. Will enjoy staying in this conversation.

an ArtsJournal blog