Delayed auditory feedback (DAF) is a technique in which someone speaks aloud, and hears their voice come back to them at a delay. In all reverberant environments, we hear the echoes of our voice when we speak (as well as the echoes of other sounds) and a lot of the early perceptual processing of sound involves cleaning the acoustic signal up to get rid of these. Around 50ms, however, the delay becomes both noticeable and disruptive to speaking aloud, and at 200ms it becomes maximally disruptive for most people. If you extend the delay over 200mns, it typically stops being so problematic.

People respond to DAF in a variety of ways: some people become very dysfluent and start to ‘stammer’, some people make their voices very flat in pitch and rhythm: some people utterly bellow. Furthermore some people can remain quiet fluent (though their voice will sound affected in some way) while others simply cannot speak. Notable, many people with a developmental dysfluency (e.g. stammering) find it easier to speak under DAF (or other altered auditory environments, such as choral speaking, altered pitch feedback, even a noisy room)

The question of why DAF affects speech is still being unraveled: the maximum disruption of 200ms is about the duration of a normal syllable in speech, and Pete Howell at UCL has argued that this means you are trying to start a new syllable just as you heard the start of last syllable you said: in other words, it makes it hard to speak because it disrupts the timing of normal speech. We have found that the dominant cortical activation associated with DAF is associated with bilateral posterior auditory fields (see below). The brain areas associated with detecting the delays and in compensating for the delays in speech output are in parts of the brain that have been linked to sensory-motor representations and processes of sound and speech (Warren link). Current work in our lab is trying to establish the neural basis of why some people do better (and worse) than others on this task.

Example of peak in right auditory cortex to increasing amount of DAF

DAF has been suggested as a way of deterring unpleasant chanting at football games though this has serious limitations. People were interested in a more sinister application of DAF as a way of jamming the speech of an individual using a directional microphone  (e.g. when “some people tend to lengthen their turns or deliberately disrupt other people when it is their turn … rather than achieve more fruitful discussions”). Need we worry about this? It’s clearly at least technically feasible, but remember (1) not everyone is as affected by DAF and everyone else (and some people are dramatically improved by the technique). There are DAF apps which you can use to find out what DAF is like, if you you want to know more about how you'd react (e.g. http://itunes.apple.com/gb/app/daf-assistant/id309496166?mt=8). Also, (2) consider other technology that can come to your aid – wear noise cancelling headphones, and turn them on if you suspect long range DAF might be used on you. 

Prof. Susan Greenfield has been back in the news complaining that spending time on twitter means “it's five hours not giving someone a hug, not feeling the sun on your face, not feeling the wind in your hair, not having an ongoing relationship in three dimensions."

(the debate and the ensuing twitter rinsing is nicely summarised in @neurobonker’s blog http://neurobonkers.com/2012/02/25/twitter-vs-susan-greenfield/) . I know we all know that she’s promoting a book, that she’s not on twitter nor does she research it or publish in the area. It may therefore come as absolutely no surprise that there are some really essential points that she’s missing about human communication. I work in human communication, and I need to get this off my chest.

 

Humans use language overwhelmingly to have conversations. We learn to speak and to understand speech in conversation, and wherever people share a language; they use it to have conversations. It’s been suggested that humans use conversation to replace the social grooming seen in other primates (http://www.springerlink.com/content/m56626627867t471/), and conversation does fulfill many of the same roles of social grooming. We can maintain social links based on very few shared words. And it’s really important to us: there are examples from the deaf literature of people walking for a day to see someone who also signs their language, just so they can have a chat. In this light, much of what goes on, on Twitter, is people using a slightly different medium to do what they’ll do any way they can, which is to converse, to talk to others. For humans, conversation is an end it itself.

 

Clearly, then,  face-to-face conversation may be the default mode for conversations but it’s not the only way we have conversations. Phone calls, texts, emails, chat rooms, product reviews, blog comments, facebook status comments: these area all ways that we replicate conversations across a variety of modes and with different degrees of privacy. There are some amazing aspects of spoken conversation, which display our virtuoso spoken language and conversational skills. For example, the temporal aspects of conversation (spoken or signed) are incredibly well controlled: people rarely speak over one another and when one person stops speaking the next person takes their turn with extreme precision (even if you are talking on the phone with a total stranger) (see figure for some conversations, spoken and signed, as well as some principles of conversation).

 

Figure: Principles of conversation

Conversation is like a dance, only instead of dancing in synchrony, we take turns. This is possible partly because when we talk to other people we align a lot of our behavior with them: we align our breathing, we align our pronunciations and our speech rhythms. This alignment with the other people we are talking is probably why we are limited in the number of people we can talk to – it’s been estimated that the maximum number of people who can share a conversation is five. Above that, some people don’t get to talk at all, or sub-conversations start to break out. Irritate your friends by running them in different group size conditions to see this in action.

 

Interestingly, if you free people from the demands of having to organize all the stuff in a face-to-face conversations that is concerned with the turn-taking negotiations, then conversations can really flourish. People can leap from one conversation to the next, and back and forth, when the time line is fast and busy, as it is for many people on Twitter (or chat rooms etc.). In these forums, you are seeing the human ability to converse set free from those demands of the “three dimensional” reality. And people have even found a way to have hugs. There, there. *hugs*.

 

People use conversations and language for all sorts of reasons – from gossip to arguments. That’s normal for conversation and it’s normal for twitter (I entirely accept that some people are ruder in electronic formats than they might be in face-to-face conversation). But Greenfield’s monomania on the idea that Twitter exists solely for people to tell other people what they are doing is as silly as suggesting that we only talk to each other explain what we are currently doing. I use twitter to share things that I find interesting, to promote work from my lab, and to join conversations that I find interesting. And that’s before I’ve stalked impressionists whose brains I want to scan. Or used it as a teaching. Or asked people for help. Or……well, you get my drift.

 

I’m glad to note that the irony of Greenfield making this claim on *television*,  of all conversation-killing places, doesn’t seem to be lost anyone. 

You can download the .pptx of my talk from the British Voice Association meeting on 15th Jan 2011 here: (http://www.divshare.com/download/16605105-96c ) (pdf here: http://www.divshare.com/download/16605144-d24 ). There are links in it to youtube clips and relevant open access papers etc.. Do email me for more details if necessary (sophie.scott@ucl.ac.uk)

Why were so many people (myself included) convinced that the dog in the “Jesus Christ in Richmond Park” (http://www.youtube.com/watch?v=3GRSbr0EYYU) video was called Benton? There are at least two reasons: first, the recording is noisy, due to the wind and the noise of people mobbing around on the grass: this masks the initial /f/ in ‘fenton’ which is itself a noisy sound which is not voiced (i.e. the vocal folds are not vibrating): perhaps if the dog were called Venton we would be less confused. The spectrograms below show the spectro-temporal profile of a me saying “Fenton Benton” and examples from the video of a ‘Fenton!’ a ‘FENTON’ and an ‘Oh, Jesus Christ’. Time is on the x axis and frequency is on the y axis: where there is more energy, the colours are brighter. In the examples of me speaking, you can see that there is a lot of noisy energy at the onset of Fenton which reflects the /f/ sound: the start of Benton is more abrupt and less noisy. The noise of the /f/ is harder to see in the video recordings, because of the level of background noise. The background noise shows up as a continuous and energy level across all frequencies in the recording.

The second reason is that the name Fenton may be less common than Benton (the latter generates about twice as many google hits, for example). We don’t have many examples of the speech of Fenton’s owner (all he says in the video is ‘Fenton. FENTON! And “Jesus CHRIST”) so we aren’t getting much context from what he says or how he speaks to help us with this. Notably, in the spectrogram below, when the owner says “oh Jesus Christ” I ‘hear’ a /t/ at the end of Christ, although you can see that the noisy /t/ burst is as obscured at the end of Christ as the /f/ is at the start of Fenton. If we expect to hear something, we can be disposed to hear it.

Public engagement is an important part of our lives as researchers, scientists and academics, and I have found it to be highly rewarding, both personally and scientifically. But what happens when the interaction is less positive, and can we learn anything from this?

An academic friend of mine was approached three years ago to provide background information for a famous author, in a ‘friend-of-a-friend’ capacity. The author was writing a book about a particular psychiatric condition. The author didn’t record the session (just took notes), and was quite insistent on certain lines of questions, pertaining to the frequency of the condition in ‘real life’ and examples of how it may manifest. 

When the book came out, my friend was surprised to learn that this one, short meeting is described as a series of meetings and phonecalls - these simply never took place. Maybe this is artistic license, in the sense that the author aimed to create a more interesting narrative from the single event. More worrying, however, was the tone of the comments ascribed to my friend. In the context of a discussion of a psychiatric condition, there is no way any professional would make public comments in a flippant way, but flippant and inappropriate is how my friend’s comments appear in the book. This is more uncomfortable ground and leaves them with the sensation not only of words being put in their mouth, but unprofessional and unfair words at that. Things get even worse: a ‘quote’ is used that suggests my friend found a particular patient sexually attractive, which takes us from flippant and inappropriate comments to total misquotations generating the impression that my friend is highly unprofessional. Remember, this is a real patient being discussed, and my friend is a real researcher. Thus, developing an interesting narrative thread has put a real person at real risk, both personally and professionally. By chance, my friend got wind of a pre-print version of the book, and the author was asked to at least remove the name of the hospital where the patient is based (to make both patient and researcher less identifiable). The author was happy to do this, and it was removed from the UK print version, though it’s still there in my Kindle version. 

What could have been different? I’m sure the famous writer meant no harm or maleficent intent towards my friend. Maybe we just need to take it on the chin that if someone is interviewing us for a popular book, it’s all up for grabs and we won’t get much say in how we are portrayed. Maybe we should expect some artistic license in how meetings and introductions are presented. Maybe, also, we could be more careful in establishing what the purpose of a meeting is and how the information we provide will be used. To academics, what we understand by a background research meeting does not necessarily mean that we realise we will be quoted (let alone quoted inaccurately). Maybe we could insist that the interview is recorded, or that we get to see some approval of the finished product. And just maybe, the people who come to us asking for help with their writing could bear in mind that they are sometimes taking away a lot when they use our names and our research to carry a story.

Although Sophie Scott has been an esteemed Professor of Cognitive Neuroscience for almost 5 years, she hasn't yet managed to give an inaugural lecture, which is traditional after election to a Chair at UCL. We're delighted to announce that this lecture will now take place on Monday 11th July 2011. 

Just a quick plug for one more event that we are running with Reeps One as part of the Vocal Invention event at the Norwich Arts Centre. We're giving a 1-hour talk, with performance from Reeps One and science from Dr Carolyn McGettigan - entrance is £3 and the talk takes place in the Norwich Arts Centre Cafe Bar at 7pm.

Roll up roll up! After a very successful evening at the ICA we are pleased to announce a second event with UK Champion Beatboxer Reeps One as part of Brain Awareness Week 2011.

For those of you who can’t make it to the event at the ICA (‘Noises in Your Head’):

 

We have recently been working with the UK Champion beatboxer, Reeps One, to explore what happens in his head when he performs.

 

Below is a video of real-time movements of the articulators (lips, tongue etc) during beatboxing – this was collected using a special rapid MRI (magnetic resonance imaging) sequence that captured 3-4 images per second. The sound you can hear in the background in the noise made by the scanner as it acquires the images. Most people are usually impressed by the tongue in these scans, not just for its size but for how nimbly it jumps around inside the oral cavity.

 

We also ran an fMRI (functional MRI), this time to look at neural activity during beatboxing. We asked Reeps One and a control participant (a non-expert) to either Beatbox, Count or Rest in the MRI scanner. The video below illustrates how this worked – you can watch what Reeps could see in the scanner and listen to how he performed the task.

 

The images below show the comparison of activations for Beatbox > Count for Reeps (in yellow), and the other showing the same contrast in the control participant (Magenta). Beatboxing and Counting are both reasonably complex articulations but both participants showed  significantly more activation for Beatboxing than Counting in motor cortex (shown on the sagittal (side-on) images). In Reeps' case, this is probably because he articulates at an amazing rate when beatboxing, and also uses an extraordinary range of articulations. In the case of our ‘novice beatboxer’, she was being asked to do something difficult and unlearned, hence the increased motor effort. Also interesting is the neat bilateral activation in lobule VI of the cerebellum for Reeps. This part of the cerebellum is strongly implicated in a network of brain areas involved in sensorimotor processing.

 

In terms of directly comparing the two brains, we can't say very much for now as we have only one professional and one control. However, it is interesting that Reeps shows very strong, focal activation of primary motor cortex, somatosensory cortex and cerebellum, while our control recruits additional cortical areas including pre-SMA (pre supplementary motor area) in the midline of the brain, which is implicated in the selection/planning of articulations. She also activated additional, dorsal parts of somatosensory cortex. This could reflect her non-expertise in performing the task - she is doing more work to plan her beatboxing articulations and pay attention to the positioning of her articulators, while we might see Reeps engaging a much more overlearned motoric strategy. We'd need more subjects to see if this pattern holds between pros and novices.

 

In the temporal lobe, both subjects show activation for Beatbox > Count in a posterior medial site on the planum temporale (on the top of the temporal lobe). This site has been implicated in a dorsal route for auditory processing, termed the 'how' pathway because it has been linked to the conversion of sound to action/articulation in auditory perception, and to verbal working memory. In contrast, listening to meaningful input like words and sentences typically engages an anterior-going pathway – the 'what' stream - which proceeds ventrally along the temporal lobe and is implicated in the processing of sounds for meaning. The motorically demanding task of beatboxing (when compared with counting) seems to have preferentially engaged the dorsal 'how' pathway in both our participants so far.

If you are interested in hearing more about this, look out for more events during Brain Awareness Week (March 14th-20th 2011).