Mapping 3D-shaping processing in the macaque monkey: a combined fMRI and single-cell study

By combining fMRI and electrophysiological recordings within the same monkeys we tried to obtain more insight in the parietal cortical areas sensitive to depth structure. In particular, awake monkey fMRI revealed that anterior IPS is activated more for curved surfaces compared to flat surfaces at different disparities. However, using the same stimuli set and fMRI experimental design, two separate activations in the IPS were observed in humans by Georgieva et al. (2009), namely DIPSA and DIPSM. This difference between monkeys and humans could be due to low spatial resolution in the monkey fMRI study, or genuine interspecies differences in the organization of anterior IPS. We targeted the activated fMRI voxels within this area with single-cell recordings and observed two patches of higher-order disparity selective neurons, one more anterior and one more posterior in AIP. These results reconcile previous observations of human and monkey fMRI studies concerning sensitivity to depth structure. Furthermore the results of this study show that with awake monkey fMRI the patchy distribution of the 3D structure selective neurons could not be detected, although fMRI generally localizes the neurons correctly. Accordingly, the relation between fMRI responses and neural activity remains complex and awake monkey fMRI in combination with electrophysiology can be a valuable technique for gaining more insight. Thanks to technological improvements, i.e. higher signal-to-noise ratio, higher spatial resolution and solid behavioral control we could define a more extensive stereo network (activated more to curved surfaces compared to flat surfaces at different disparities) compared to previous studies. New (besides the anterior IPS and F5a) robust activations were found in caudal IPS, PIT and more consistent and extended activations were observed in AIT. We reversibly inactivated area CIP, one of the early stages of the dorsal stream in the stereo network, while performing the same fMRI experiment to investigate the causal influence it exerts on other components of the stereo network. We found causal contributions of CIP on the depth structure related activations in the anterior IPS and surprisingly also to those in AIT in the ventral visual stream. In addition, CIP inactivation reduces the performance of the monkeys on a depth structure categorization task. We argue that the CIP inactivation effects in anterior IPS and AIT are caused indirectly since electrical microstimulation of area CIP and PIP did not activate anterior IPS nor AIT. These results designate a key role to area CIP in processing depth structure information for both the dorsal and the ventral visual stream.Table of Contents List of abbreviations iii Acknowledgements v Chapter 1: General Introduction 1 1.1 Processing of visual information 1 1.2 Seeing in depth 3 1.2.1 Binocular disparity 3 1.3 Flow of 3D information from early visual areas to higher-order visual areas 5 1.3.1 Evidence coming from single cell studies 5 1.3.2 Evidence coming from fMRI 11 1.3.3 Anatomical connections between ventral and dorsal stream 14 1.4 Methodology 15 1.4.1 Function magnetic resonance imaging 15 1.4.2 Relation between single-cell studies and fMRI 16 1.4.3 Reversible inactivation 19 Objectives 21 Chapter 2: The Relationship between functional magnetic resonance imaging and single-cell selectivity in the intra-parietal sulcus 25 2.1 Abstract 25 2.2 Introduction 26 2.3 Materials and Methods 28 2.3.1 Subjects, surgery and training procedures 28 2.3.2 Stimuli 31 2.3.3 fMRI scanning procedures 32 2.3.4 Single-cell recording procedures 34 2.3.5 Data analysis 38 2.4 Results 42 2.4.1 fMRI results: curvature x disparity interaction effect in the IPS 42 2.4.2 Assessment of higher-order disparity-selectivity 45 2.4.3 Patches of higher-order disparity-selective neurons 47 2.4.4 Population spiking responses 54 2.4.5 Local field potential responses 58 2.4.6 Localizer experiments 60 2.5 Discussion 65 2.6 Supplemental figures 72 Chapter 3: Functional interactions between dorsal and ventral visual stream during three-dimensional objects vision 75 3.1 Abstract 75 3.2 Introduction 76 3.3 Methods 77 3.3.1 Subjects, surgery and stimuli 77 3.3.2 Training in depth-structure categorization 79 3.3.3 Scanning procedures 80 3.3.4 Reversible inactivation procedures 82 3.3.5 Electrical Microstimulation 83 3.3.6 Data Analysis 84 3.4 Results 88 3.4.1 The network of depth structure-sensitive cortical areas 89 3.4.2 The effect of reversible inactivation of CIP during fMRI (ipsilateral results) 93 3.4.3 The effect of reversible inactivation of area CIP on depth structure categorization 101 3.4.4 CIP microstimulation during fMRI activates a distinct network of cortical areas 103 3.5 Discussion 106 3.6 Supplemental figures 115 Chapter 4: General Discussion 121 4.1 Summary of findings 121 4.2 Limitations 123 4.2.1 Limitations of fMRI 123 4.2.2 Limitations of electrophysiology 125 4.2.3 Limitations of the stimuli 125 4.2.4 What did we inactivate? 126 4.2.5 Near future experiments 127 4.2.6 Future perspectives 128 4.2.7 Relation to human studies 129 4.2.8 Conclusion 130 Summary 133 Samenvatting 135 Reference List 137 Curriculum Vitae 153nrpages: 156status: publishe