Representation Of Spatial Information In The Macaque Posterior Of The Superior Temporal Polysensory Area(STPp)

The Superior Temporal Polysensory area(STP)of rhesus monkeys is located at the end of the visual pathway,receiving projections from both dorsal(motion signals)and ventral pathways(object recognition),as well as inputs from other non-visual sensory information.Thus,it is an area where multisensory signals highly converges.Until now,few studies have been carried on this brain region,and the neuron response characteristics and functional roles are unclear.In the current study,we focused on the posterior portion of STP(STPp)and explores whether this area encodedself-motion information and/or represents three-dimensional(3D)objects.To explore whether STPp is involved in self-motion cognition,we used a virtual reality system to accurately match visual and vestibular stimuli to simulate self-motion.By the extracellular recording of STPp neuron activity,we found that 6.3% of STPp neurons were significantly tuned to visual and vestibular orientation(p<0.05,One-Way Anova).The proportion of neurons with similar visual and vestibular preferences("congruent neurons")is close to the neurons with opposite visual and vestibular preferences("opposite neurons").In STPp,18.3% of neurons tuned only to the vestibular signal,smaller than that of neurons tuned only to visual(30.4%).The strength of vestibular tuning was also weaker than that of visual.Compared to other multisensory brain regions responding to both visual and vestibular signals,the vestibular and visual tuning strength of STPp was less than the dorsal medial superior temporal area(MSTd),the ventral intra-parietal area(VIP),and the visual posterior sylvian area(VPS).In addition,the latency of visual response of STPp was significantly later than MSTd,but the latency of vestibular response was significantly faster than MSTd,suggesting that the vestibular signals related to self-motion in STPp might not come from the input of the upstream visual brain region MSTd.When vestibular and visual stimuli were applied simultaneously,the firing rate increased than that in single cue conditions but less than the sum of the single cue condition.However,adding vestibular signals did not affect the visual preference.These results indicate that STPp is dominated by visual signals,and a small number of neurons carry self-motion signals,but the contribution might be small.To explore whether STPp is involved in 3D spatial information processing,we used 3D planar stimuli based on binocular disparity,in which 3D planar orientation is defined by Slant and Tilt and depth is defined by Mean Disparity.By the extracellular recording of neuron activity in STPp,we found that the proportion of neurons tuned to Mean Disparity in STPp at Slant 0°(28.3%)was lower than that at Slant 30° and 60°(63.2%,51.3%),indicating that most neurons in STPp are sensitive to the relative disparity.In addition,90.2% of neurons did not change the preference for Mean Disparity under different Slant angles,suggesting that Slant and Mean Disparity signals are encoded independently in STPp.The proportion of Tilt tuned cells in STPp under Slant 60°(29.6%)was higher than that under Slant 30°(13.8%),and 84.21% of neurons’ preference change for Tilt between Slant 30? and Slant 60° was less than 45°,suggesting that Slant and Tilt are also encoded independently in STPp.In addition,about 10% of STPp neurons are Tilt-dependent disparity tuned neurons(TDD),which means that neurons’ preference for Mean Disparity reverses in the opposite Tilt direction.60% of TDD neurons’ Tilt preference was also reversed with Mean Disparity,suggesting that STPp may be involved in more complex object recognition(such as second-order surfaces).Finally,to examine whether STPp can stably encode 3D planar information,the motion information is added to the stimuli,which means the points move along the plane from the inside out at the neurons’ preferred speed.Compared to the 3D planar stimuli without motion information,the proportion of neurons tuned to Tilt and Mean Disparity increased in STPp under motion conditions,suggesting that motion information can improve the ability of STPp to encode plane information.However,no significant change was observed either in the 75.8% of neurons’ Tilt preference or 81.8%of neurons’ Mean Disparity preference,suggesting that STPp can stably recognize 3D objects.In summary,STPp carries a small amount of self-motion information and has an obvious representation of 3D object plane information,but the functional role of STPp still needs to be explored.