Dogs can sense weak thermal radiation

The dog rhinarium (naked and often moist skin on the nose-tip) is prominent and richly innervated, suggesting a sensory function. Compared to nose-tips of herbivorous artio- and perissodactyla, carnivoran rhinaria are considerably colder. We hypothesized that this coldness makes the dog rhinarium particularly sensitive to radiating heat. We trained three dogs to distinguish between two distant objects based on radiating heat; the neutral object was about ambient temperature, the warm object was about the same surface temperature as a furry mammal. In addition, we employed functional magnetic resonance imaging on 13 awake dogs, comparing the responses to heat stimuli of about the same temperatures as in the behavioural experiment. The warm stimulus elicited increased neural response in the left somatosensory association cortex. Our results demonstrate a hitherto undiscovered sensory modality in a carnivoran species.


2
After extensive praising and rewarding, the dog was lifted off the bed and all left the scanning room for a cca.5-minute break.

B. Experiment
During the break between the pre-conditioning phase and the measurements, E prepared the 'stimulus-presenting' device for the experiment by filling up the glass cuboid with warm tap water. The temperature of the water was adjusted to the temperature in the scanning room and was controlled by a thermographic camera.
Before and after each experimental run, E took thermographic measurements of the dogs' nose and the warm stimulus' temperature. Since these measurements could not be taken in the scanning room where the actual experiment took place due to technical reasons, they were taken in the waiting room. However, the difference between the ambient temperature of the waiting room and the scanning room was on average only 1.2 °C (SD=0.6°C). The temperature of the scanning room was measured by a built-in thermometer and was displayed in the computer room.
The device was mounted on the scanning bed again. The dog was lifted on the bed and was asked to put its nose in the required position. During the experiment (and pre-conditioning), E, T and O had assigned positions in the scanning room. E and O were at the front end of the scanner, E was sitting on a chair, invisible to the dog, while O was standing, looking at the dog throughout the experiment. T was standing at the far end of the scanner, having both the dog and the Op in the computer room in her line of sight. Since both O and T had a reliable view on the dog, they could detect and end the measurement in case the dog was moving. E presented the stimuli by operating the device according to a pre-determined order.
Instructions on the order and the timing of door openings were recorded on an audio file and were played back to E on a headphone.

3
At the end of the test, the Op signaled T who then signaled the others in the scanning room.
After extensive praising and rewarding, the dog was lifted off the bed and all left the scanning room.

fMRI experiment-Data acquisition
The MRI measurements were performed using a Philips Ingenia 3.0 T whole body MR system.
We used a Philips SENSE Flex Medium coil. The EPI-BOLD fMRI time series were obtained from 30 transverse slices, covering the whole brain. The spatial resolution was 2×2×2 mm, with a 0.5 mm slice gap. A single-shot gradient-echo planar sequence was used (ascending slice order; acquisition matrix 96×57; TR = 3200 ms, including 2000 ms acquisition and 1200 ms silent gap; TE = 30 ms; flip angle = 90°). All runs included 75 volume acquisitions. As anatomical reference, a standard T1-weighted three-dimensional scan was collected (turbo-field echo TFE) sequence; 180 slices; 1×1×1 mm spatial resolution). Threshold of motion was set to 3 mm translation or more than 3° rotation in any direction. Runs exceeding the threshold were discarded and repeated. In case of unsuccessful runs, the dogs were placed in the scanner at a maximum of 3 times in a row. 5 Figure S1. Device generating the stimuli in the behavioral experiment. a, Cross-section through a custom-crafted panel used to generate warm and neutral stimuli for behavioral training and testing. The heated or the insulated surface was turned toward the dog to generate a warm or neutral, respectively, stimulus. b, Photograph of the stimulation apparatus with two panels on sliding mounts. The warm side is pushed back, exposing the food reward underneath.
There was an equivalent reward also under the neutral side, but the sliding mechanism was blocked, such that the dog could not get access. For changing the stimuli from warm to neutral and vice versa, the panels stayed in place, but were turned around, meaning that there always was a warm and a neutral surface on both sides of the stimulation apparatus. Both panels were in the forward positions (closed) when the dog made its choice (see Fig. S2a). voltage DC never changed and did therefore not indicate on which side the warm surface of the panel was facing the dog. b, Experimenter's view. The blind behind which the dog had to wait while a trial was set up is opened and the dog waits to be called into the arena. From the frame holding the roller blind, about 1 m above the head of the dog, a fan was blowing approx. 45 degrees downward toward the presenting apparatus. Note the divider that separates the warm and neutral sides, forcing the dog to make its choice from a distance of at least 1.6 m.