ICMPC Day Two – Amusia

OK, show time! The end of Day Two was marked by the Amusia session… and by the Amusia session, I mean my friend and colleague Dr Fang Liu and I giving our talks to a relatively small but lovely audience. I was especially happy to see some of the leaders in our field were there though including Gottried Schlaug and Simone Dalla Bella. It did nothing for my nerves, mind you, but it was nice to see them in any case!

Fang started the session by presenting some of her fantastic studies of intonation perception and production in amusia, which have been published in Brain since we applied for this conference over a year ago. Fang’s work builds on an inconsistency regarding amusics’ abilities with language based stimuli. Originally studies seemed to suggest that amusic individuals had no difficulty detecting and reproducing pitch changes in speech (intonation), but more recently a study by Patel indicated that some amusics might have difficulties. Fang had speculated that perhaps the answer lay in the degree of pitch changes used in different experiments. Amusics don’t have difficulty with large pitch changes in music, so why should they have difficulty with the same size leaps in language? Better to look at small pitch changes, so more subtle intonation inflections. Do they have a difficulty perceiving or producing these?

Fang carefully constructed her stimuli to manipulate small rises or falls in the end of sentences, thereby creating statements or questions. The amusics had to try to identify the sentences as a statement or a question and have a go at reproducing the sounds. Fang measured their accuracy, their production for later speech analysis, and movements of their vocal apparatus using a clever measuring instrument that fits around the glottis. She also measured their pitch perception thresholds for the detection of a pitch change and the discrimination of pitch direction.

She found that amusics did have difficulties with both perception and production of speech intonation compared to controls, when the changes were subtle. But this did not reflect a lack of control of the vocal apparatus. She also showed that deficits in pitch detection, and more severely in the discrimination of pitch direction. Fang’s work is incredibly precise, well-controlled, thoughtfully carried out, plus her statistical analysis and presentation is second to none. I feel very privileged to have worked with her for two years and I wish her all the best at UCL where she is now based.

Then it was my turn. Fang is not an easy act to follow at all, but thankfully she had already introduced a lot of background for me, making my job easier. I began by reiterating the phenotype of amusia, including my ‘what it is like to be amusia’ slide, which I hope brings it home to people what it is actually like to have this condition. Then I presented a little background on the MBEA test to show how amusia is diagnosed. Following this I introduced background about pitch perception studies of amusia, those which reported a difficulty in fine-grained pitch perception. Then I introduced more recent studies that have implicated a concomitant pitch memory problem, one which cannot entirely be explained by the pitch perception deficit. Studies have shown amusics have difficulty holding on to a single tone in memory, either in conditions of interference or time delay, and that their memory for 4 or 5 tone sequences was poorer than that of controls. My research question was how many tones can amusics hold in their memory, when you control for pitch perception difficulties? Furthermore, if they do have a pitch memory difficulty (as we suspected they would) how far does this problem spread into other memory systems?

So I conducted equivalent adaptive tracking span tests for digits and tones, and took measures of working memory. We controlled for perception ability by taking the same types of threshold measures as Fang and making the minimum tone gap in our pitch memory task 4 times greater than the average threshold, and twice as great as that of the worst amusic. We found that controls and amusics could remember around 7 digits on average, so no difference in ability there. Nor was there a difference in overall group working memory. However, whilst controls could remember about 7 tones, amusics could only get to 4. So here we have evidence for a pitch memory problem that cannot be explained by perceptual difficulties or difficulties higher up in the memory system. The implications of this are that there is likely to be some pitch-specific processing going on in memory, which is somehow weaker in amusia and consequently that amusia is more than a fine-grained pitch perception difficulty.

I got quite a few questions, which is always a lovely end to a talk.

i) A lady asked me why there was a range in amusics tone memory scores. I think she expected them to be all lying at the bottom of the graph. I pointed out that the range of scores for amusics in the tone memory task was equivalent to that seen in their digit span and that seen in both the controls’ span tasks. So they have a normal range of ability within the task, but at a much lower level. I also pointed out that amusia is a neurodevelopmental condition, so ability is likely to be different across individuals who demonstrate a similar main phenotype, as in dyslexia.

ii) Simone asked if we had the MBEA memory subtest scores to correlate with our findings. A good question! Sadly we don’t use that subtest in our diagnosis, so I can’t do that analysis, but it was a nice idea.

iii) Gottfried asked if I have tried to partial out the effects of digit memory from our tone memory results to see the size of the effect that remained. Another very good idea! I might try that when I get home.

iv) And finally someone pointed out that pitch perception difficulties could lead to memory difficulties by nature of starving the memory system of training input during development, a point I agree with entirely. We still don’t know what causes the memory problem or exactly where the memory problem lies within the pitch memory system (i.e. is it encoding, rehearsal, retrieval…?). All interested questions for the future.

I really enjoyed out little amusia session, and a few people were kind enough to approach me afterwards and say that they liked it too. All their comments were greatly appreciated and I look forward to presenting our new findings in the visual domain soon (hopefully with Fang again :-))


  • Bjørn Petersen

    Hey Vicky
    I was at your talk and tried to contact you, but you were a bit busy afterwards. You made a very good presentation.
    I am doing a phd-project on musical ear training with cochlear implant users – dhildren and adults. They are in a way amusic too, though rather tecnologically induced than neurologically.
    Anyway, yóu mentioned som tests that were freeware! I was not fast enough to write daomn the webpage. Could you possibly post them?
    Best regards

  • vicky

    Hi, nice to hear from you. Sorry I missed you, it is one of the consequences of having a lot of nice people came to talk to me afterwards. Sure, there are lots of working memory tests available here thanks to the brilliant chaps at Randall Engle’s Lab: http://www.psychology.gatech.edu/renglelab/Tasks.htm. Very best wishes for your PhD. I look forward to hearing about the findings.