Stereomotion processing in the non-human primate brain

Abstract

The cortical network that processes disparity-defined motion-in-depth (i.e. cyclopean stereomotion) was characterised with functional magnetic resonance imaging in two awake, behaving macaques. The experimental protocol was similar to previous human neuroimaging studies. We contrasted the responses to dynamic random-dot patterns that continuously changed their binocular disparity over time with those to a control condition that shared the same properties, except that the temporal frames were shuffled. A whole-brain voxel-wise analysis revealed that in all four cortical hemispheres, three areas showed consistent sensitivity to cyclopean stereomotion.Two of them were localised respectively in the lower bank of the superior temporal sulcus (CSM STS ) and on the neighbouring infero-temporal gyrus (CSM ITG ). The third area was situated in the posterior parietal cortex (CSM PPC ). Additional ROIs-based analyses within retinotopic areas defined in both animals indicated weaker but significant responses to cyclopean stereomotion within the MT cluster (most notably in areas MSTv and FST). Altogether, our results are in agreement with previous findings in both human and macaque and suggest that the cortical networks that process cyclopean stereomotion is relatively well preserved between the two primate species.In the present study, we used fMRI recordings in macaque to determine the cortical regions that have specific responses to stereomotion based on changing disparity over time (CDOT). We used an experimental protocol that was directly adapted from previous human studies (Likova and Tyler;Rokers et al., 2009; Kaestner et al., 2019). Materials and Methods SubjectsTwo female rhesus macaques (age: 15-17 years; weight: 5.35-6.15 kg) were involved in the study. Animal housing, handling, and all experimental protocols (surgery, behavioural training, and MRI recordings) followed the guidelines of the European Union legislation (2010/63/UE) and of the French Ministry of Agriculture (décret 2013-118). All projects were approved by a local ethics committee (CNREEA code: C2EA -14) and received authorisation from the French Ministry of Research (MP/03/34/10/09). Details about the macaques' surgical preparation and behavioural training are provided elsewhere (Cottereau et al., 2017). Data AvailabilityData and analysis code will be made available after acceptance of the paper on dedicated platforms (PRIME-DE and OSF: https://osf.io/yxrsv/). Experimental designOur stimuli were derived from those of previous fMRI studies that investigated how cyclopean stereomotion (motion-in-depth based on CDOT) is processed in humans Pre-processing of the raw functional dataPre-processing and volume-based analyses were carried with SPM12 in the Matlab environment (MathWorks®). Only runs with central gaze fixation above 85% were kept for further analysis. In total, we kept 43 and 60 runs for both macaques, respectively. The 4 first volumes of each run were discarded (dummy scans) to Mapping, 29(4), 411-421.

Decoding neural responses to motion-in-depth using EEG

“…One possibility is that a mechanism agnostic to differences in CD and IOVD cue properties, or possibly fuses both velocity and disparity signals together (Movshon & Newsome, 1996; Ponce et al, 2008). One candidate cortical location for this general stereomotion processing mechanism is the region in or around the human MT+ complex which includes cells that are known to be sensitive to both lateral motion defined from a variety of cues, as well as 3D-motion (Czuba et al, 2014; Héjja-Brichard et al, 2020; Huk, 2012; Kaestner et al, 2019; Likova & Tyler, 2007; Rokers et al, 2009; Sanada & DeAngelis, 2014). The temporal integration of 3D-motion signals has been shown to occur across hundreds of milliseconds, from ~150 to 1000 ms post-stimulus, with sensitivity to 3D-motion increasing across this integration period (Katz et al, 2015) and the behavioural integration of 3Dmotion signals across time is known to be relatively slow (when compared to lateral motion) (Brooks & Stone, 2006; Brooks & Stone, 2004; Harris et al, 2008; Harris & Watamaniuk, 1995, 1996; Huk, 2012; Norcia & Tyler, 1984; Richards, 1972).…”

“…Recent functional magnetic resonance neuroimaging (fMRI) and psychophysical studies suggest that 3D-motion might be encoded downstream; areas in and around MT have been shown to respond to MID (Likova & Tyler, 2007;Rokers et al, 2009) and MT neurons are selective for 3Dmotion directions (Czuba et al, 2011;Joo et al, 2016;Rokers et al, 2009Rokers et al, , 2011. Likewise, fMRI studies have identified that these cues both activate motion selective area MT (or its human homologue) and nearby regions, while electrophysiological studies in macaque have shown that CD and IOVD stimuli both drive responses in macaque MT (Czuba et al, 2014;Héjja-Brichard et al, 2020;Likova & Tyler, 2007;Rokers et al, 2009;Sanada & DeAngelis, 2014). However, psychophysical measurements of CD and IOVD suggest that these mechanisms adapt independently (Joo et al, 2016;Sakano et al, 2012), have different speed sensitivities (Shioiri et al, 2008;Wardle & Alais, 2013), and engage systems with different spatial resolutions (Czuba et al, 2010).…”

Decoding Neural Responses to Motion-in-Depth Using EEG

“…One possibility is the existence of a mechanism that is agnostic to differences in CD and IOVD cue properties, or possibly fuses both velocity and disparity signals together (Movshon and Newsome, 1996;Ponce et al, 2008). One candidate cortical location for this general stereomotion processing mechanism is the region in or around the human MT+ complex which includes cells that are known to be sensitive to both lateral motion defined from a variety of cues, as well as 3D-motion (Likova and Tyler, 2007;Rokers et al, 2009;Huk, 2012;Czuba et al, 2014;Sanada and DeAngelis, 2014;Kaestner et al, 2019;Héjja-Brichard et al, 2020). The temporal integration of 3D-motion signals has been shown to occur across hundreds of milliseconds, from ∼150 to 1000 ms post-stimulus, with sensitivity to 3Dmotion increasing across this integration period (Katz et al, 2015) and the behavioral integration of 3D-motion signals across time is known to be relatively slow (when compared to lateral motion) (Richards, 1972;Norcia and Tyler, 1984;Watamaniuk, 1995, 1996;Stone, 2004, 2006;Harris et al, 2008;Huk, 2012).…”

“…Recent functional magnetic resonance neuroimaging (fMRI) and psychophysical studies suggest that 3D-motion might be encoded downstream; areas in and around MT have been shown to respond to MID (Likova and Tyler, 2007;Rokers et al, 2009) and MT neurons are selective for 3D-motion directions (Rokers et al, 2009(Rokers et al, , 2011Czuba et al, 2011;Joo et al, 2016). Likewise, fMRI studies have identified that these cues both activate motion selective area MT (or its human homolog) and nearby regions, while electrophysiological studies in macaque have shown that CD and IOVD stimuli both drive responses in macaque MT (Likova and Tyler, 2007;Rokers et al, 2009;Czuba et al, 2014;Sanada and DeAngelis, 2014;Héjja-Brichard et al, 2020). However, psychophysical measurements of CD and IOVD suggest that these mechanisms adapt independently (Sakano et al, 2012;Joo et al, 2016), have different speed sensitivities (Shioiri et al, 2008;Wardle and Alais, 2013), and engage systems with different spatial resolutions (Czuba et al, 2010).…”