deadpopstar.com

I haven’t been posting a lot of neuroscience content lately, despite my aim to steer this blog in that direction. After spending 10 hours a day on neuroscience, writing about it isn’t usually the first thing that comes to mind upon returning home. But I’m going to try! And I’m going to cheat by writing about things that happen to be in my field.

So, without further ado: When hearing loss is caused by some fault of the machinery that converts sound energy into neural impulses within the cochlea, a spiral shaped structure within the inner ear, a cochlear implant (CI) is usually the only option in restoring hearing. It works by electrically activating remaining neurons, using an electrode array placed in the scala tympani, one of three fluid filled chambers of the cochlea (the coiled structure, below).

Now, the cochlea has what is known as a frequency-place code. Different regions of the cochlea are tuned to particular frequencies, with high frequency sounds down at the base, and low frequency sounds at the tip. So, by stimulating different regions, you can create the perception of different pitches. A CI, in essence, pulls out the dominant frequencies of the received auditory signal, and stimulates the corresponding region of the cochlea. In the very best patients, this gives near perfect speech perception — which is a fascinating phenomenon given how much information is stripped out by the CI.

There are a few limitations of the CI that researchers are trying to address. The most oft-stated are poor speech recognition in noisy environments (such as a busy restaurant), and poor music perception. It’s suggested that if you can deliver more detailed pitch information, it will allow the brain to filter out some noise, or recognize complex melodies.

Given what I’ve just said, you might imagine that if you can stimulate more regions of the cochlea, you can represent more pitches. Unfortunately, this is a difficult proposition. Unlike the normal mechanical fine tuning of the cochlea, the electrical stimulation from a CI spreads over a broad area, activating neurons that represent a range of frequencies. If you put electrodes too close together, there will be so much overlap in the groups of activated neurons that the brain won’t be able to tell the difference. In the case of speech perception, it’s been found that stimulating anything more than about eight electrodes yields no improvement (a typical human CI has about 22 electrodes to choose from).

Part of the problem is that the electrode array lies in the fluid of the scala tympani, often with a layers of protein and cell encapsulation, at some distance from the neurons it’s trying to stimulate. Therefore, you’re unlikely to get greater frequency selectivity with increased numbers of electrodes, unless you can somehow get the electrodes nice and close to those neurons.

Stained cochlea cross-section. The * shows where the electrode array was, with a tissue capsule left behind (black arrow). The blue arrow shows where the auditory neuron cell bodies are located, with their axons forming the auditory nerve down the centre (the modiolus).

There’s been a couple of approaches on that one. The first is trying to skewer the cochlea with an electrode array right down the middle of the cochlea, in the auditory nerve itself. Two problems with this are that it is (a) invasive, and (b) now difficult to determine which neurons correspond to which frequencies. But you do get lower thresholds (i.e. less power is required, which is very exciting to engineers like myself), and access to a broader frequency range (current clinical electrode arrays can’t reach the upper turns of the cochlea). Middlebrooks & Snyder (2008) recently provided a review of this approach. Another approach has been to insert an electrode array into the scala tympani, but get it to hug the wall against the modiolus, as close as possible to the neurons (there are also efforts afoot to get the neurons to grow onto the electrode array, but that’s another story).

Contour Electrode Array (Cochlear Ltd.)

The latter has been in clinical use for a while, but a clear benefit to auditory performance has yet to fully emerge, suggesting it isn’t the silver bullet some were hoping for. The means by which the curved electrode array is inserted is clever, but by no means advanced. During surgery, a stylette sits inside the electrode array, keeping it straight. As the array is inserted into the cochlea, the stylette is incrementally removed, allowing the tip to curl around the cochlea, eventually hugging the wall of the modiolus once fully removed. In short, it relies on the skill of the surgeon, knowing where the electrode tip is, and how it is situated. Pull the stylette out too slowly, and the electrode could bang against the wall of the cochlea instead of following the turns of the cochlea. Too quickly, and the tip could wrap back on itself.

I’m overstating the danger a bit; the array design has been optimized to reduce the risk of such hiccups. But wouldn’t it be nice to know where the electrode array is in the cochlea, how it is oriented, and if it is about the crash into something important, rather than just going by feel? That’s what Wise and colleagues (2008) set out to do at The University of Michigan. They built a pre-curved silicon electrode array with sensors along it’s length, and at its tip. Combined with some onboard processing, this permits the shape of the electrode array, and forces exerted upon it, to be monitored during insertion.

Photo showing the smart electrode array, with the inset showing a stimulation electrode (black circle) and position/contact sensors (the orange squiggly things).

The authors hope that this system will eventually be integrated with an automated insertion tool, allowing electrode arrays to be positioned with precision and repeatability. And, being silicon based, more electrode sites can be integrated than clinical (hand-made) electrode arrays, like the Contour. Though, as I’ve mentioned, it’s unclear whether this will provide added benefit to most patients. Lack of cochlear trauma, and preservation of any existing hearing, however, certainly provides benefit. In those with residual hearing, the best outcomes are achieved when a cochlear implant is combined with a hearing aid (usually used for low frequencies, where the electrode array can’t reach) — an approach called bimodal hearing.

Wong, Skoe, Russo, Dees & Kraus (2007) Nature Neuroscience. 10 (4): 420-422.
Abstract: Music and speech are very cognitively demanding auditory phenomena generally attributed to cortical rather than subcortical circuitry. We examined brainstem encoding of linguistic pitch and found that musicians show more robust and faithful encoding compared with nonmusicians. These results not only implicate a common subcortical manifestation for two presumed cortical functions, but also a possible reciprocity of corticofugal speech and music tuning, providing neurophysiological explanations for musicians’ higher language-learning ability.

I’ve been meaning to go over this paper for some time, after being first mentioned a few months ago. It’s a fascinating study for two reasons. Firstly, it shows that experience-related (music, in this case) plasticity in the auditory pathway is not necessarily specific to one class of stimuli. The musical experience gained by subjects in this study seems to have helped with their ability to perceive linguistic patterns. The more traditional view is that there is some sort of strict dichotomy between music and language, the most facile example being that music is right-brained and language is left-brained. The second reason why I find this study fascinating is that this effect was observed in the brainstem. That is, before the cortex — which is usually depicted as the highly flexible and dynamic organ responsible for learning.

The authors took a sample of musically-trained individuals and compared their Frequency Following Response (FFR) with that of controls when listening to pitch-differentiated Mandarin words. Mandarin is known as a tone-language, in that much of the information is conveyed in the pitch of the word. In this case, a /mi/ sound was used, which has three different meanings depending on whether a rising, dipping or level inflection is used.

The FFR, as you might imagine, reflects tracking of pitch in scalp-recorded potential, and is thought to originate from the Inferior Colliculus (a structure just before the thalamus, which serves as a relay station for most sensory information). Best demonstrated by the representative figures below:

Ah, representative figures. The grey line shows the change in frequency (pitch) of the dipping /mi/ sound, and the yellow line shows it being tracked by the FFR. Note that the musician was able to track the pitch of this (previously unfamiliar) Mandarin word, while his or her counterpart in the control group could not. The authors interpreted this as evidence that the musician brain more faithfully encodes pitch information. Representatives cases aside, Wong et al also demonstrated pitch tracking differences between groups, with the effect strongest in those who started musical training early and stuck with it for a long time. It’s worth noting that even though this response is observed prior to the cortex, it is probably still driving plasticity via top-down feedback.

The major implication of this research, hinted at in the abstract above, is that musical training can drive language development - something well worth considering when developing K-12 curriculum.

The 2007 International Congress of Neuroscience ran in Melbourne from the 12th to the 17th of July. It’s not every day that one of the world’s biggest neuro conferences is in your hometown! Yes, I only just got around to blogging it. Please forgive me if the details are a little fuzzy.

The first session I attended was a plenary by Prof. Mandyam Srinivasan, a neuroethologist who studies bee behaviour at ANU (and, more recently, at UQ). This was probably the best neuroscience lecture I had ever seen - I’d love to be able to give such charming presentations. Mandyam has amassed a substantial collection of high impact papers detailing the bee vision system, in particular, the concept of optic flow.

Because bee eyes are so close together, they don’t get much in the way of 3D information via parallax. However, they are able to infer distances by measuring how much of the world rushes by; the closer you are, the faster the surface appears to move. Think driving down a highway - trees by the road rush by, whilst hills in the distance remain static. This can be demonstrated by training bees to fly down a tunnel, with the striped walls attached to a sort of conveyor belt. By moving the walls at different speeds, you can mess with the bee’s flight down the narrow tunnel. Humans use these cues too, by the way.

He also talked about some work on decoding bee dances, extending research that earned Karl von Frisch the Nobel Prize in 1973. Srinivasan’s group has demonstrated that distances to food sources communicated by these dances are measured with respect to optic flow, such that the bees can be tricked into giving misleading dances if they travel down the tunnel with striped conveyor belt walls to get their food reward. It also seems that when bees traverse over low contrast terrain, such as calm open water, they perceive distances as being smaller (link to PLoS Biology paper).

This is shown in the photo above, in the path heading south west . The feeding station (white dots) was gradually moved away from the hive, over land, then water, then land again on the island.

Distance is communicated to other bees in the hive using a “waggle dance“. Briefly, the bee does a little loop, and then a zig-zag type motion (”waggle”), the duration of which is directly proportional to the distance to the food source (a sucrose solution). As shown above, the distance communicated is smaller when over low-contrast terrain, like open water. This is reflected in significance testing of curve fit slopes for each segment. This suggests that the bees are using optic flow to measure distance, as opposed to, say, energy consumption. An interesting tale lies in how one might conduct an experiment to test this alternate hypothesis, but I’ll save that one for another day.

The final talk of the visual motion processing symposium, given by Prof. Dario Floreano, director of the Laboratory of Intelligent Systems at École Polytechnique Fédérale de Lausanne in Zurich, also covered optic flow. It was quite different to the other talks, and certainly appealed to my inner engineer. Dario uses optic flow to get tiny autonomous flying robots (below) to navigate and avoid obstacles.

The planes fly in a room dubbed the holodeck, where different images, of varying contrast, can be projected onto the walls. Despite being sold as having heady military applications, Dario pointed out that this was mostly just for fun. My kind of project! That tiny 10 gram plane includes two cameras, two gyroscopes, an accelerometer, an anemometer, a bluetooth radio, and a microcontroller!

Finally, I invented the following award.

Award for Best Item of Exhibitor Swag
Not being a clinical conference, there weren’t any picnic baskets or trips to Tahiti on offer. There was, however, this…

It’s a reflective slap band thingy you put around your trouser leg to stop it getting caught in your bike chain! Get it? Neuroscientists are dorks who ride bikes with their pants tucked into their socks! I love it!

The Neurocritic reports on a paper in Australasian Psychiatry which finally reveals why Sydneysiders live in overpriced real estate, are woefully inept at sport, and are unable to land an NHMRC grant. (Link, Follow up)

I find avocado slicers creepy. They remind me of the leucatome, a surgical implement used to perform lobtomies; a retractable loop of wire is inserted into the brain, typically through the eye socket, and twisted — rendering all your psychiatric problems a thing of the past!

(RH picture from Nobel Foundation).

Elsewhere: Tenuously related but awesome - The Neurophilosopher: An illustrated history of trepanation

Cochlear implants (CIs) treat sensorineural hearing loss by electrically stimulating the auditory nerve with a electrode array placed in the scala tympani, one of three fluid filled chambers of the cochlea. It won’t work, however, if the neural pathway is severed between the inner ear and brain. In these cases, the only remaining option is an auditory brainstem implant (ABI), which stimulates the brainstem at the cochlear nucleus where the auditory nerve makes its first connection.

This is not without difficulty. The CI works in large part due to the orderly tonotopic organisation of the cochlea - frequencies are represented along the length of the coiled organ, with high frequencies at the base, and low frequencies at the tip, or apex. While tonotopicity is preserved throughout the auditory system, it’s not nearly so accessible as in the cochlea. ABIs also involve neurosurgery about the brainstem, a region not to be idly meddled with, given that it also governs vital functions like arousal and respiration.

The culmination of these and other difficulties is that ABIs don’t work; at least, not to the extent that CIs permit speech perception in many individuals. This is not unlike early CI models, which were originally intended to cue lip reading. But, as the technology has matured, far better outcomes have been achieved - especially in those implanted at a very young age. This is thought to be due to a greater capacity for neural reorganisation, or plasticity; the ability to adapt to the impoverished auditory information provided by the CI. In contrast, ABIs, are largely restricted to providing environmental auditory cues, not speech perception.

So, with that rather extended introduction; The Age reports that three-year-old Jorja Steele received an auditory brainstem implant earlier this month , which is a very rare undertaking in a child. The outcome will be fascinating - one would hope that as in young CI recipients, Jorja’s brain will learn to adapt to the coarse stimuli provided by the implant, and make sense of it.

Article Links: High hopes for implant to penetrate Jorja’s silent world, Small hole opens Jorja’s mind to a sound future

(Second Photo: Penny Stephens, Jason South)

A study to be published in Nature Neuroscience next month by Wong and colleagues suggests that early musical training enhances the capacity of the brainstem to track pitch changes in tone-based languages such as Mandarin, ostensibly observed through evoked responses¹. It’s a particularly interesting finding in that plasticity related to such high level cognition is usually associated with with the cerebral cortex; while the brainstem does extract features for speech processing, plastic changes were only thought to be of significance with profound changes in auditory experience, such as in congenital deafness.

I’m still curious about their actual methodology, their sample, discussions of sensitive periods and their implications for clinical interventions and so fourth, but that’ll have to wait a couple of weeks until the actual unmolested paper becomes available.

Press Release (via Omni Brain)

(more…)

The mechanisms that combine multiple sensory modalities, beginning the formation of a complete percept, were explored by Kayser and colleagues - and it turns out that integration may occur earlier in the processing hierarchy than first thought.

J Neurosci2007; 27(8):1824.

Merging the information from different senses is essential for successful interaction with real-life situations. Indeed, sensory integration can reduce perceptual ambiguity, speed reactions,or change the qualitative sensory experience. It is widely held that integration occurs at later processing stages and mostly in higher association cortices; however, recent studies suggest that sensory convergence can occur in primary sensory cortex. A good model for early convergence proved to be the auditory cortex, which can be modulated by visual and tactile stimulation; however, given the large number and small size of auditory fields, neither human imaging nor microelectrode recordings have systematically identified which fields are susceptible to multisensory influences. To reconcile findings from human imaging with anatomical knowledge from nonhuman primates, we exploited high-resolution imaging (functional magnetic resonance imaging) of the macaque monkey to study the modulation of auditory processing by visual stimulation. Using a functional parcellation of auditory cortex, we localized modulations to individual fields. Our results demonstrate that both primary (core) and nonprimary (belt) auditory fields can be activated by the mere presentation of visual scenes. Audiovisual convergence was restricted to caudal fields [prominently the core field (primary auditory cortex) and belt fields (caudomedial field, caudolateral field, and mediomedial field)] and continued in the auditory parabelt and the superior temporal sulcus. The same fields exhibited enhancement of auditory activation by visual stimulation and showed stronger enhancement for less effective stimuli, two characteristics of sensory integration. Together, these findings reveal multisensory modulation of auditory processing prominently in caudal fields but also at the lowest stages of auditory cortical processing.

Link (via Science Daily)

Shelley describes how an unfortunate chop stick through the eye facillitated the collection and subsequent culture of neural stem cells from the prefrontal subcortex.

A Chinese woman was admitted to Huashan Hospital in Shanghai, with a chopstick in her brain (!)—specifically the inferior prefrontal subcortex. The chopstick was removed by a Dr. Zu, who took the opportunity to culture the brain tissue that came out with the chopstick. [...]

The cultured tissue thrived, and many of the resultant cells contained proteins that were characteristic of neural stem cells. In order to make sure they were really stem cells, Dr. Zu cultured cells in isolation and watch and see if it divides into daughter cells. He found that about 4% of the ‘chopstick cells’ went on to form neurospheres (a ball of daughter cells), indicating that they were stem cells.

Jumping ahead…

They transplanted cultured neural stem cells derived from 8 patients with brain injuries back into those same patients’ brains. They then asked a separate group of neurologists to blindly examine these experimental patients and compare them with un-treated control patients who also had similar injuries. The treated patients had lower disability scores (a good thing), possibly paving the way for this therapy to be used clinically. And all because of one mis-aimed chopstick!

I haven’t been able to find any peer reviewed references on this latter paragraph, and would appreciate if anyone could point me in the right direction. There’s a review by Zhu and colleagues here (PDF), and a mention by Seed magazine. China’s seemingly prodigous status in stem cell research is interesting in itself.

(Pretty confocal photomicrograph from here.)

Cochlear implants are able to restore language reception in many individuals, but they lack the temporal and spectral resolution to restore the full experience of music.

I could hear the drums of Bolero just fine. But the other instruments were flat and dull. The flutes and soprano saxophones sounded as though someone had clapped pillows over them. The oboes and violins had become groans. It was like walking color-blind through a Paul Klee exhibit. I played Bolero again and again, hoping that practice would bring it, too, back to life. It didn’t.

Michael Chorost details his quest to seek out the latest implant speech processor software, to try and restore his pre-deafness experience of Ravel’s Bolero. Link