For those of you who followed us last year, you may have noticed one small difference in how we’re recording heart rate – we’ve rolled out our brand-spanking new “Marque II” tags for the 2014 field season! (You can have a look at the “Marque I” tags from last year here or below).
We used our last field season to gather as much data as possible – not just on the seals themselves, but also the on the equipment we were using. When you spend eight hours a day watching seals for seven weeks, you definitely want to get your “bang for your buck”!
One of the biggest hurdles we’ve had to overcome is the presence of “artefacts” in our traces. As Amy summarised rather nicely last week, one of the main reasons we’re monitoring heart rates is to understand how individual seals deal with stress. Since we need to record every single heartbeat to do this, when a handful of beats are missing or spurious, it deteriorates the quality of our traces and limits what inferences we can make about how stressed an individual is.
Artefacts ranged from easily corrected blips to prolonged stretches of peaks and troughs or “flats”. Some artefacts had a clear origin – if a seal rolls over on its back, the sensor on the tag is definitely going to struggle to transmit through 200 kg of seal! – whereas others were occurring without any obvious signs, when individuals were resting.
Where were these artefacts coming from? Were they an after-effect of particularly active behaviours (e.g. aggression with a neighbour)? Was the local topography deflecting the signal from the tag (e.g. gullies or hollows)? Were they more likely to happen when a female decided to go for a dip in wallow? These are all questions we asked, because although artefacts can be treated with fancy correction algorithms – as the saying goes – prevention is the best cure.
A lot of the analyses we’ve done on data from last year has focused on pinpointing potential sources of artefacts in our system. As it turns out, one of the main issues was with the design of the tags themselves – or more specifically, how we sealed the electrode gel (also known by the highly-scientific term, “goo”) within the tag, which is needed to optimize conductance between the skin and the electrodes.
We noticed artefacts became more common the longer it had been since “goo-ing” – a pattern that cropped up across all of our females. After a little pondering, we realised that the design of the Marque I tag was to blame. Since the electrodes were essentially “fused” in the middle (like a saddle), there was a lot of room for goo to move as it pleased under the tag – away from where it’s needed, under the electrodes – or even worse, leak out and evaporate.
Cue some creative modification, and the Marque II was born. The key difference was we used electrodes that were “articulated”, allowing us to form a complete seal around each electrode. This makes it miles better at retaining the goo for prolonged periods of time (and saving this year’s research team the exquisite agony of watching goo ooze out of a freshly-deployed Marque I tag).
Glancing at the traces we’re getting for this field season, we have reason to be cautiously optimistic: the artefacts don’t seem to follow the same patterns we observed with the Marque I tags. Whether our modifications have made a real difference will have to wait until the last tag of the season comes off and we trudge wearily back to our desks in the lab.
No doubt the Marque II won’t be the last modification we make (a Marque III is already in the back of our minds!), but it’s one crucial step in the right direction. Until then, all we can do is hunker down against the cold Scottish autumn and watch our seals with our fingers crossed!