Canine sense of quantity: evidence for numerical ratio-dependent activation in parietotemporal cortex

Abstract

47The approximate number system, which supports the rapid estimation of quantity, emerges early 48 in human development and is widespread across species. Neural evidence from both human and 49 non-human primates suggests the parietal cortex as a primary locus of numerical estimation, but 50 it is unclear whether the numerical competencies observed across non-primate species are 51 subserved by similar neural mechanisms. Moreover, because studies with non-human animals 52 typically involve extensive training, little is known about the spontaneous numerical capacities 53 of non-human animals. To address these questions, we examined the neural underpinnings of 54 number perception using awake canine functional magnetic resonance imaging. Dogs passively 55 viewed dot arrays that varied in ratio and, critically, received no task-relevant training or 56 exposure prior to testing. We found evidence of ratio-dependent activation, which is a key 57 feature of the approximate number system, in canine parietotemporal cortex in the majority of 58 dogs tested. This finding is suggestive of a neural mechanism for quantity perception that has 59 been conserved across mammalian evolution. 60 61 Keywords: approximate number system; canine cognition; quantity discrimination; fMRI; 62 parietal cortex 63 93 of cortex for representing non-symbolic numerical quantity, then activation in this region should 94 increase as the ratio between alternating dot arrays increases, despite constant cumulative surface 95 area and variable element size [17]. That is, a number-sensitive region of cortex will exhibit 96 greater activation when the numerical values of the stimuli presented are more dissimilar (e.g., 2 97 vs. 10 dots) than when numerical value is constant (e.g., 6 vs. 6 dots), consistent with Weber's 98 law [18]. 99 100 Methods 101 Participants 102 103 Eleven awake, unrestrained dogs (see Table 1 for demographic information), were 104 scanned in a Siemens 3T Trio MRI scanner. Prior to testing, all dogs completed a training 105 program to be desensitized to the scanner environment through behavior shaping and positive 106 reinforcement [19]. All dogs had previously participated in fMRI studies while viewing stimuli 107 on a projection screen but had no prior training on numerical discrimination. 108 109 Table 1. Dogs' demographic information Stimuli were 75 dot arrays comprised of light gray dots on a black background (800 x 130 800 px). For each numerosity used (2, 4, 6, 8, and 10), stimuli varied in cumulative area (i.e., the 131 total gray on each image). For each numerosity, cumulative area was either 10, 20, or 30% of the 132 total stimulus. For each numerosity, 15 unique stimuli were used (5 stimuli per cumulative area 133 value). In each stimulus, individual dot size varied up to 30%. Dot location varied randomly. 134Critically, these controls minimize the influence of non-numerical properties, in order to ensure 135 that the results can be attributed to changes in numerical value [20, 21]. In accordance with 136 current esti...