Image from flickr user ~db~ under a Creative Commons CC BY-NC-ND 2.0 Licence
When we describe the behaviour or movement of a human or any animal, we usually categorise behaviours into discrete units – for example running, climbing or throwing. But we often take for granted that these kinds of movement are a result of unlimited possibilities of potential movements. But what if the brains of animals, including us, can’t in fact produce any possible combination of movement, but that our brains select from a collection of set movement types?
Historically, it hasn’t been possible to experimentally determine which of these two possible scenarios is correct, in part because we haven’t had fast enough cameras to distinguish the two – that is we have lacked temporal resolution.
In a study published online on January 4th in Current Biology, researchers at the Champalimaud Centre for the Unknown in Lisbon used high speed video tracking of individual larval Zebrafish (the favourite fish for a lot of research labs) to investigate this question. One reason for using these Zebrafish larvae is that they are previously been found to have quite discrete movement behaviours and that there are 3 distinct levels of Zebrafish locomotion – a “half-beat” (essentially, a flap of the tail), a sequence of half-beats called a bout, and a sequence of bouts strung together to achieve something – move a certain direction at a particular speed, or explore the area around the fish.
To attempt to get a better understanding of whether these behaviours truly are discrete and to identify how many movement types Zebrafish really use, the authors of this paper used high speed video tracking (at 700 frames per second). To try to catch a wide variety of behaviour types, swimming of the larvae was recorded when fish were alone in a tank as well as when multiple fish were present, the fish were also given a variety of stimuli, such as light or food, to see how there movement was affected. The researchers then used an algorithm they developed to automatically categorise the behaviour of fish into distinct movement types.
What they found was that they could categorise the repertoire of Zebrafish movements into 13 different movement units (a press release for the paper described these as “Lego bricks” of movement). These included very general movement patterns such as “slow swim 1” and more specialise ones such as “short capture swim” used when hunting (Zebrafish eat microscopic invertebrates called Rotifers). They also found that particular sequences of these bouts are more likely than others in response to specific stimuli.
This is all quite cool and the paper supports the notion that Zebrafish larvae have a discrete set of behavioural “bouts” which are combined to achieve particular goals. But, to a large extent this paper has just used modern technologies to investigate zebrafish larval movement in a better and less biased way. They have identified some novel movement types. Studies in adult Zebrafish have found possible discrete modes of swimming, but for other movements, data suggest a continuum of movement possibilities. Mammals obviously have a much wider range of potential movements and behaviours, making this kind of behavioural analysis less useful. Mice do show different gaits at different speeds, but whether this is truly discrete behavioural types isn’t clear. So while I think the way this study was carried out was pretty neat, I think it should really be filed in the “interesting but not widely applicable” folder.