The following Neural Mechanisms On-line webinar “Hyperalignment: Modeling shared data encoded in idiosyncratic cortical topographies” will likely be delivered by James V. Haxby on June third. See beneath for particulars in regards to the free speak and how you can be part of.
Hyperalignment:
Modeling shared data encoded in idiosyncratic cortical topographies
James V. Haxby (Dartmouth School)
h 14-16 Greenwhich Imply Time / 16-18 CEST (check your local time here)
| Be a part of the assembly by way of Webex (the hyperlink will likely be activated 10 minutes earlier than the speak) |
Summary. Info that’s shared throughout brains is encoded in idiosyncratic fine-scale useful topographies. Hyperalignment captures shared data by projecting sample vectors for neural responses and connectivities into a typical, high-dimensional data house, reasonably than by aligning topographies in a canonical anatomical house. Particular person transformation matrices undertaking data from particular person anatomical areas into the widespread mannequin data house, preserving the geometry of pairwise dissimilarities between sample vectors, and mannequin cortical topography as mixtures of overlapping, individual-specific topographic foundation capabilities, reasonably than as contiguous useful areas. The elemental property of mind operate that’s preserved throughout brains is data content material, reasonably than the useful properties of native options that assist that content material. On this Perspective, we current the conceptual framework that motivates hyperalignment, its computational underpinnings for joint modeling of a typical data house and idiosyncratic cortical topographies, and focus on implications for understanding the construction of cortical useful structure.
You could find the paper here.
Sensible data:
Movies of prior talks may be discovered on the Neural Mechanisms YouTube channel.
