Coupling of total hemoglobin concentration,oxygenation, and neural activity in rat somatosensory cortex

Recent advances in brain imaging techniques, including functional magnetic resonance imaging (fMRI), offer great promise for noninvasive mapping of brain function. However, the indirect nature of the imaging signals to the underlying neural activity limits the interpretation of the resulting maps. The present report represents the first systematic study with sufficient statistical power to quantitatively characterize the relationship between changes in blood oxygen content and the neural spiking and synaptic activity. Using two-dimensional optical measurements of hemodynamic signals, simultaneous recordings of neural activity, and an event-related stimulus paradigm, we demonstrate that (1) there is a strongly nonlinear relationship between electrophysiological measures of neuronal activity and the hemodynamic response, (2) the hemodynamic response continues to grow beyond the saturation of electrical activity, and (3) the initial increase in deoxyhemoglobin that precedes an increase in blood volume is counterbalanced by an equal initial decrease in oxyhemoglobin.     

New paradigm for optical imaging: temporally encoded maps of intrinsic signal

We present a new technique for acquiring and analyzing intrinsic signal optical images of brain activity, using continuous stimulus presentation and data acquisition. The main idea is to present a temporally periodic stimulus and to analyze the component of the response at the stimulus frequency. Advantages of the new technique include the removal of heart, respiration, and vasomotor artifacts, a dramatic increase in spatial resolution, and a 30-fold or greater reduction in acquisition time. We also present a novel approach to localizing instantaneous neuronal responses using time-reversed stimuli that is widely applicable to brain imaging. To demonstrate the power of the technique, we present high-resolution retinotopic maps of five visual areas in mouse cortex and orientation maps in cat visual cortex.     

High-resolution fMRI maps of cortical activation in nonhuman primates: correlation with intrinsic signal optical images

One of the most widely used functional brain mapping tools is blood oxygen level-dependent (BOLD) functional magnetic resonance imaging (fMRI). This method has contributed to new understandings of the functional roles of different areas in the human brain. However, its ability to map cerebral cortex at high spatial (submillimeter) resolution is still unknown. Other methods such as single- and multiunit electrophysiology and intrinsic signal optical imaging have revealed submillimeter resolution of sensory topography and cortical columnar activations. However, they are limited either by spatial scale (electrophysiology characterizes only local groups of neurons) or by the inability to monitor deep structures in the brain (i.e., cortical regions buried in sulci or subcortical structures). A method that could monitor all regions of the brain at high spatial resolution would be ideal. This capacity would open the doors to investigating, for example, how networks of cerebral cortical columns relate to or produce behavior. In this article we demonstrate that, without benefit of contrast agents, at a magnetic field strength of 9.4 tesla, BOLD fMRI can reveal millimeter-sized topographic maps of digit representation in the somatosensory cortex of the anesthetized squirrel monkey. Furthermore, by mapping the "funneling illusion," it is possible to detect even submillimeter shifts in activation in the cortex. Our data suggest that at high magnetic field strength, the positive BOLD signal can be used to reveal high spatial resolution maps of brain activity, a finding that weakens previous notions about the ultimate spatial specificity of the positive BOLD signal.     

Covariations between pupil diameter and supplementary eye field activity suggest a role in cognitive effort implementation

In both human and nonhuman primates (NHP), the medial prefrontal region, defined as the supplementary eye field (SEF), can indirectly influence behavior selection through modulation of the primary selection process in the oculomotor structures. To perform this oculomotor control, SEF integrates multiple cognitive signals such as attention, memory, reward, and error. As changes in pupil responses can assess these cognitive efforts, a better understanding of the precise dynamics by which pupil diameter and medial prefrontal cortex activity interact requires thorough investigations before, during, and after changes in pupil diameter. We tested whether SEF activity is related to pupil dynamics during a mixed pro/antisaccade oculomotor task in 2 macaque monkeys. We used functional ultrasound (fUS) imaging to examine temporal changes in brain activity at the 0.1-s time scale and 0.1-mm spatial resolution concerning behavioral performance and pupil dynamics. By combining the pupil signals and real-time imaging of NHP during cognitive tasks, we were able to infer localized cerebral blood volume (CBV) responses within a restricted part of the dorsomedial prefrontal cortex, referred to as the SEF, an area in which antisaccade preparation activity is also recorded. Inversely, SEF neurovascular activity measured by fUS imaging was found to be a robust predictor of specific variations in pupil diameter over short and long-time scales. Furthermore, we directly manipulated pupil diameter and CBV in the SEF using reward modulations. These results bring a novel understanding of the physiological links between pupil and SEF, but it also raises questions about the role of anterior cingulate cortex (ACC), as CBV variations in the ACC seems to be negligible compared to CBV variations in the SEF.

Ultrafast functional imaging reveals short- and long-term covariations between pupil diameter and activity in the Supplementary Eye Field (SEF) of awake behaving non-human primates, yielding a novel understanding of the physiological links between the pupil and SEF.     

Retinotopic maps, spatial tuning, and locations of human visual areas in surface coordinates characterized with multifocal and blocked FMRI designs

The localization of visual areas in the human cortex is typically based on mapping the retinotopic organization with functional magnetic resonance imaging (fMRI). The most common approach is to encode the response phase for a slowly moving visual stimulus and to present the result on an individual's reconstructed cortical surface. The main aims of this study were to develop complementary general linear model (GLM)-based retinotopic mapping methods and to characterize the inter-individual variability of the visual area positions on the cortical surface. We studied 15 subjects with two methods: a 24-region multifocal checkerboard stimulus and a blocked presentation of object stimuli at different visual field locations. The retinotopic maps were based on weighted averaging of the GLM parameter estimates for the stimulus regions. In addition to localizing visual areas, both methods could be used to localize multiple retinotopic regions-of-interest. The two methods yielded consistent retinotopic maps in the visual areas V1, V2, V3, hV4, and V3AB. In the higher-level areas IPS0, VO1, LO1, LO2, TO1, and TO2, retinotopy could only be mapped with the blocked stimulus presentation. The gradual widening of spatial tuning and an increase in the responses to stimuli in the ipsilateral visual field along the hierarchy of visual areas likely reflected the increase in the average receptive field size. Finally, after registration to Freesurfer's surface-based atlas of the human cerebral cortex, we calculated the mean and variability of the visual area positions in the spherical surface-based coordinate system and generated probability maps of the visual areas on the average cortical surface. The inter-individual variability in the area locations decreased when the midpoints were calculated along the spherical cortical surface compared with volumetric coordinates. These results can facilitate both analysis of individual functional anatomy and comparisons of visual cortex topology across studies.     

Imaging odor-evoked activities in the mouse olfactory bulb using optical reflectance and autofluorescence signals

In the brain, sensory stimulation activates distributed populations of neurons among functional modules which participate to the coding of the stimulus. Functional optical imaging techniques are advantageous to visualize the activation of these modules in sensory cortices with high spatial resolution. In this context, endogenous optical signals that arise from molecular mechanisms linked to neuroenergetics are valuable sources of contrast to record spatial maps of sensory stimuli over wide fields in the rodent brain. Here, we present two techniques based on changes of endogenous optical properties of the brain tissue during activation. First the intrinsic optical signals (IOS) are produced by a local alteration in red light reflectance due to: (i) absorption by changes in blood oxygenation level and blood volume (ii) photon scattering. The use of in vivo IOS to record spatial maps started in the mid 1980's with the observation of optical maps of whisker barrels in the rat and the orientation columns in the cat visual cortex(1). IOS imaging of the surface of the rodent main olfactory bulb (OB) in response to odorants was later demonstrated by Larry Katz's group(2). The second approach relies on flavoprotein autofluorescence signals (FAS) due to changes in the redox state of these mitochondrial metabolic intermediates. More precisely, the technique is based on the green fluorescence due to oxidized state of flavoproteins when the tissue is excited with blue light. Although such signals were probably among the first fluorescent molecules recorded for the study of brain activity by the pioneer studies of Britton Chances and colleagues(3), it was not until recently that they have been used for mapping of brain activation in vivo. FAS imaging was first applied to the somatosensory cortex in rodents in response to hindpaw stimulation by Katsuei Shibuki's group(4). The olfactory system is of central importance for the survival of the vast majority of living species because it allows efficient detection and identification of chemical substances in the environment (food, predators). The OB is the first relay of olfactory information processing in the brain. It receives afferent projections from the olfactory primary sensory neurons that detect volatile odorant molecules. Each sensory neuron expresses only one type of odorant receptor and neurons carrying the same type of receptor send their nerve processes to the same well-defined microregions of ?100μm(3) constituted of discrete neuropil, the olfactory glomerulus (Fig. 1). In the last decade, IOS imaging has fostered the functional exploration of the OB(5, 6, 7) which has become one of the most studied sensory structures. The mapping of OB activity with FAS imaging has not been performed yet. Here, we show the successive steps of an efficient protocol for IOS and FAS imaging to map odor-evoked activities in the mouse OB.     

The effects of minimum map unit in coral reefs maps generated from high resolution side scan sonar mosaics

Changes in coral habitats from benthic maps at two minimum mapping units (4 and 400 m²) were examined across 600 ha at the Puerto Rico insular shelf. The maps were produced by visually interpreting high resolution side scan sonar seafloor mosaics. Comparisons were performed by using published spatial indices based on patch size and abundance. At the fine scale, small coral patches were dominant and had more complex shapes, while coarse scale maps lost information on small or rare habitats, but included new mixed habitat categories that helped conserve estimates of total habitat area. Variations in spatial indices occurred across spatial scales, but trends were not necessarily predictable. For instance, habitat diversity increased at the coarse scale maps, but habitat richness remained unchanged. Differences were related as much to the peculiarities of the abundance, shape and arrangement of habitats at the two sites as to changes in grain size.     

A conceptual framework for the spatial analysis of functional trait diversity

Functional trait diversity is a popular tool in modern ecology, mainly used to infer assembly processes and ecosystem functioning. Patterns of functional trait diversity are shaped by ecological processes such as environmental filtering, species interactions and dispersal that are inherently spatial, and different processes may operate at different spatial scales. Adding a spatial dimension to the analysis of functional trait diversity may thus increase our ability to infer community assembly processes and to predict change in assembly processes following disturbance or land‐use change. Richness, evenness and divergence of functional traits are commonly used indices of functional trait diversity that are known to respond differently to large‐scale filters related to environmental heterogeneity and dispersal and fine‐scale filters related to species interactions (competition). Recent developments in spatial statistics make it possible to separately quantify large‐scale patterns (variation in local means) and fine‐scale patterns (variation around local means) by decomposing overall spatial autocorrelation quantified by Moran's coefficient into its positive and negative components using Moran eigenvector maps (MEM). We thus propose to identify the spatial signature of multiple ecological processes that are potentially acting at different spatial scales by contrasting positive and negative components of spatial autocorrelation for each of the three indices of functional trait diversity. We illustrate this approach with a case study from riparian plant communities, where we test the effects of disturbance on spatial patterns of functional trait diversity. The fine‐scale pattern of all three indices was increased in the disturbed versus control habitat, suggesting an increase in local scale competition and an overall increase in unexplained variance in the post‐disturbance versus control community. Further research using simulation modeling should focus on establishing the proposed link between community assembly rules and spatial patterns of functional trait diversity to maximize our ability to infer multiple processes from spatial community structure.     

Assessment of single-vessel cerebral blood velocity by phase contrast fMRI

Current approaches to high-field functional MRI (fMRI) provide 2 means to map hemodynamics at the level of single vessels in the brain. One is through changes in deoxyhemoglobin in venules, i.e., blood oxygenation level–dependent (BOLD) fMRI, while the second is through changes in arteriole diameter, i.e., cerebral blood volume (CBV) fMRI. Here, we introduce cerebral blood flow–related velocity-based fMRI, denoted CBFv-fMRI, which uses high-resolution phase contrast (PC) MRI to form velocity measurements of flow. We use CBFv-fMRI in measure changes in blood velocity in single penetrating microvessels across rat parietal cortex. In contrast to the venule-dominated BOLD and arteriole-dominated CBV fMRI signals, CBFv-fMRI is comparable from both arterioles and venules. A single fMRI platform is used to map changes in blood pO₂ (BOLD), volume (CBV), and velocity (CBFv). This combined high-resolution single-vessel fMRI mapping scheme enables vessel-specific hemodynamic mapping in animal models of normal and diseased states and further has translational potential to map vascular dementia in diseased or injured human brains with ultra–high-field fMRI.

This study presents a phase contrast-based, high field MRI-based approach for the functional mapping of cerebral blood velocity in individual cortical arterioles and venules in the rat cortex; this approach can be combined with previously established approaches to map BOLD, CBV, and blood velocity from penetrating microvessels.     

High-resolution functional magnetic resonance imaging of the animal brain

To fully understand brain function, one must look beyond the level of a single neuron. By elucidating the spatial properties of the columnar and laminar functional architectures, information regarding the neural processing in the brain can be gained. To map these fine functional structures noninvasively and repeatedly, functional magnetic resonance imaging (fMRI) can be employed. In this article the basic principles of fMRI are introduced, including specific hardware requirements and the equipment necessary for animal magnetic resonance research. Since fMRI measures a change in secondary hemodynamic responses induced by neural activity, it is critical to understand the principles and potential pitfalls of fMRI techniques. Thus, the underlying physics of conventional blood oxygenation, cerebral blood flow, and cerebral blood volume-based fMRI techniques are extensively discussed. Tissue-specific signal change is close to the site of neural activity, while signals from large vessels can be distant from the actual active site. Thus, methods to minimize large vessel contributions and to maximize tissue signals are described. The fundamental limitation of fMRI spatial resolution is the intrinsic hemodynamic response. Based on our high-resolution fMRI studies, the hemodynamic response is regulated at submillimeter functional domains and thus spatial resolution can be achieved to an order of 100 microm. Since hemodynamic responses are sluggish, it is difficult to obtain very high temporal resolution. By using an approach with multiple experiments with different stimulus conditions, temporal resolution can be improved on the order of 100 ms. With current fMRI technologies, submillimeter columnar- and laminar-specific specific functional images can be obtained from animal brains.     

Mapping behaviorally relevant neural circuits with immediate-early gene expression

Immediate-early genes have gained widespread popularity as activity markers for mapping neuronal circuits involved in specific behaviors in many different species. In situ immediate early gene detection methods provide cellular level resolution, a major benefit for mapping neuronal networks. Recent advances using fluorescence in situ hybridization also afford temporal resolution, enabling within-animal activity maps for two distinct behaviors. Moreover, use of transgenic mice with fluorescent reporter proteins driven by immediate early gene promoters is enabling repeated measurements, over long time scales, of cortical activity within the same animal. These methodological innovations, coupled with recent advances in fluorescence imaging and probe development, will enable large scale mapping of behaviorally relevant circuits with temporal and three-dimensional spatial resolution in experimental animals.     

Topological reorganization of odor representations in the olfactory bulb

Odors are initially represented in the olfactory bulb (OB) by patterns of sensory input across the array of glomeruli. Although activated glomeruli are often widely distributed, glomeruli responding to stimuli sharing molecular features tend to be loosely clustered and thus establish a fractured chemotopic map. Neuronal circuits in the OB transform glomerular patterns of sensory input into spatiotemporal patterns of output activity and thereby extract information about a stimulus. It is, however, unknown whether the chemotopic spatial organization of glomerular inputs is maintained during these computations. To explore this issue, we measured spatiotemporal patterns of odor-evoked activity across thousands of individual neurons in the zebrafish OB by temporally deconvolved two-photon Ca²⁺ imaging. Mitral cells and interneurons were distinguished by transgenic markers and exhibited different response selectivities. Shortly after response onset, activity patterns exhibited foci of activity associated with certain chemical features throughout all layers. During the subsequent few hundred milliseconds, however, MC activity was locally sparsened within the initial foci in an odor-specific manner. As a consequence, chemotopic maps disappeared and activity patterns became more informative about precise odor identity. Hence, chemotopic maps of glomerular input activity are initially transmitted to OB outputs, but not maintained during pattern processing. Nevertheless, transient chemotopic maps may support neuronal computations by establishing important synaptic interactions within the circuit. These results provide insights into the functional topology of neural activity patterns and its potential role in circuit function.     

On the origin of the functional architecture of the cortex

The basic structure of receptive fields and functional maps in primary visual cortex is established without exposure to normal sensory experience and before the onset of the critical period. How the brain wires these circuits in the early stages of development remains unknown. Possible explanations include activity-dependent mechanisms driven by spontaneous activity in the retina and thalamus, and molecular guidance orchestrating thalamo-cortical connections on a fine spatial scale. Here I propose an alternative hypothesis: the blueprint for receptive fields, feature maps, and their inter-relationships may reside in the layout of the retinal ganglion cell mosaics along with a simple statistical connectivity scheme dictating the wiring between thalamus and cortex. The model is shown to account for a number of experimental findings, including the relationship between retinotopy, orientation maps, spatial frequency maps and cytochrome oxidase patches. The theory's simplicity, explanatory and predictive power makes it a serious candidate for the origin of the functional architecture of primary visual cortex.     

Object-oriented methods for habitat mapping at multiple scales – Case studies from Northern Germany and Wye Downs, UK

This paper presents an application of object-oriented techniques for habitat classification based on remotely sensed images and ancillary data. The study reports the results of habitat mapping at multiple scales using Earth Observation (EO) data at various spatial resolutions and multi temporal acquisition dates. We investigate the role of object texture and context in classification as well as the value of integrating knowledge from ancillary data sources. Habitat maps were produced at regional and local scales in two case studies; Schleswig-Holstein, Germany and Wye Downs, United Kingdom. At the regional scale, the main task was the development of a consistent object-oriented classification scheme that is transferable to satellite images for other years. This is demonstrated for a time series of Landsat TM/ETM+ scenes. At the local scale, investigations focus on the development of appropriate object-oriented rule networks for the detailed mapping of habitats, e.g. dry grasslands and wetlands using very high resolution satellite and airborne scanner images. The results are evaluated using statistical accuracy assessment and visual comparison with traditional field-based habitat maps. Whereas the application of traditional pixel-based classification result in a pixelised (salt and pepper) representation of land cover, the object-based classification technique result in solid habitat objects allowing easy integration into a vector-GIS for further analysis. The level of detail obtained at the local scale is comparable to that achieved by visual interpretation of aerial photographs or field-based mapping and also retains spatially explicit, fine scale information such as scrub encroachment or ecotone patterns within habitats.     

A neural mechanism of hierarchical discrimination of odors in the olfactory cortex based on spatiotemporal encoding of odor information

We propose a neural mechanism for discrimination of different complex odors in the olfactory cortex based on the dynamical encoding scheme. Both constituent molecules of the odor and their mixing ratios are encoded simultaneously into a spatiotemporal activity pattern (limit cycle attractor) in the olfactory bulb [Hoshino O, Kashimori Y, Kambara T (1998) Biol Cybern 79:109–120]. We present a functional model of the olfactory cortex consisting of some dynamical mapping modules. Each dynamical map is represented by itinerancy among the limit cycle attractors. When a temporal sequence of spatial activity patterns corresponding to a complex odor is injected from the bulb to the network of the olfactory cortex, the neural activity state of each mapping module is fixed to a relevant spatial pattern injected. Recognition of an odor is accomplished by a combination of firing patterns fixed in all the mapping modules. The stronger the response strength of the component, the earlier the component is recognized. The hierarchical discrimination of an odor is made by recognizing the components in order of decreasing response strengths. Received: 28 November 1998 / Accepted in revised form: 17 December 1999     

Effects of invariant sympathetic activity on cutaneous circulatory responses to heat stress

This study investigated the role of sympathetic withdrawal on blood flow responses in cutaneous arteriovenous anastomoses (AVAs) and capillaries to direct and indirect heat stress. This was achieved by clamping sympathetic activity (SC) to the tail of anesthetized rats so that constrictor tone remained invariant during exposure of either the animal's tail (direct heating) or body (indirect heating) to a 35 degrees C environment. Flow through the AVAs in the tail was evaluated by laser-Doppler flowmetry (LDF), while capillary flow was investigated by videodensitometry measurements of blood cell velocity (CBV) in single capillaries within the subepidermal vascular plexus. Both direct and indirect heating significantly increased LDF and CBV. In comparison to blood flow responses in sham-operated control rats, the SC procedure resulted in significantly lower LDF responses to both direct and indirect heat stress. By contrast, the response of CBV was not significantly affected by SC during either mode of heating. These results indicate that the withdrawal of sympathetic constrictor tone plays a role in the response of cutaneous AVAs, but not precapillary arterioles, to direct as well as indirect heat stress. Additional studies on unanesthetized animals showed that superimposing body heating on a base of local heating elicited a further increase in LDF, suggesting that local heating does not deplete neural mediated dilatory reserve.     

Delayed striate cortical activation during spatial attention

Recordings of event-related potentials (ERPs) and event-related magnetic fields (ERMFs) were combined with functional magnetic resonance imaging (fMRI) to study visual cortical activity in humans during spatial attention. While subjects attended selectively to stimulus arrays in one visual field, fMRI revealed stimulus-related activations in the contralateral primary visual cortex and in multiple extrastriate areas. ERP and ERMF recordings showed that attention did not affect the initial evoked response at 60-90 ms poststimulus that was localized to primary cortex, but a similarly localized late response at 140-250 ms was enhanced to attended stimuli. These findings provide evidence that the primary visual cortex participates in the selective processing of attended stimuli by means of delayed feedback from higher visual-cortical areas.     

Standing waves and traveling waves distinguish two circuits in visual cortex

The visual cortex represents stimuli through the activity of neuronal populations. We measured the evolution of this activity in space and time by imaging voltage-sensitive dyes in cat area V1. Contrast-reversing stimuli elicit responses that oscillate at twice the stimulus frequency, indicating that signals originate mostly in complex cells. These responses stand clear of the noise, whose amplitude decreases as 1/frequency, and yield high-resolution maps of orientation preference and retinotopy. We first show how these maps are combined to yield the responses to focal, oriented stimuli. We then study the evolution of the oscillating activity in space and time. In the orientation domain, it is a standing wave. In the spatial domain, it is a traveling wave propagating at 0.2-0.5 m/s. These different dynamics indicate a fundamental distinction in the circuits underlying selectivity for position and orientation, two key stimulus attributes.     

Activity restriction, impaired capillary function, and the development of insulin resistance in lean primates

Insulin produces capillary recruitment in skeletal muscle through a nitric oxide (NO)-dependent mechanism. Capillary recruitment is blunted in obese and diabetic subjects and contributes to impaired glucose uptake. This study's objective was to define whether inactivity, in the absence of obesity, leads to impaired capillary recruitment and contributes to insulin resistance (IR). A comprehensive metabolic and vascular assessment was performed on 19 adult male rhesus macaques (Macaca mulatta) after sedation with ketamine and during maintenance anesthesia with isoflurane. Thirteen normal-activity (NA) and six activity-restricted (AR) primates underwent contrast-enhanced ultrasound to determine skeletal muscle capillary blood volume (CBV) during an intravenous glucose tolerance test (IVGTT) and during contractile exercise. NO bioactivity was assessed by flow-mediated vasodilation. Although there were no differences in weight, basal glucose, basal insulin, or truncal fat, AR primates were insulin resistant compared with NA primates during an IVGTT (2,225 ± 734 vs. 5,171 ± 3,431 μg·ml(-1)·min(-1), P < 0.05). Peak CBV was lower in AR compared with NA primates during IVGTT (0.06 ± 0.01 vs. 0.12 ± 0.02 ml/g, P < 0.01) and exercise (0.10 ± 0.02 vs. 0.20 ± 0.02 ml/g, P < 0.01), resulting in a lower peak skeletal muscle blood flow in both circumstances. The insulin-mediated changes in CBV correlated inversely with the degree of IR and directly with activity. Flow-mediated dilation was lower in the AR primates (4.6 ± 1.0 vs. 9.8 ± 2.3%, P = 0.01). Thus, activity restriction produces impaired skeletal muscle capillary recruitment during a carbohydrate challenge and contributes to IR in the absence of obesity. Reduced NO bioactivity may be a pathological link between inactivity and impaired capillary function.     

Spatial Structure of Above-Ground Biomass Limits Accuracy of Carbon Mapping in Rainforest but Large Scale Forest Inventories Can Help to Overcome

Precise mapping of above-ground biomass (AGB) is a major challenge for the success of REDD+ processes in tropical rainforest. The usual mapping methods are based on two hypotheses: a large and long-ranged spatial autocorrelation and a strong environment influence at the regional scale. However, there are no studies of the spatial structure of AGB at the landscapes scale to support these assumptions. We studied spatial variation in AGB at various scales using two large forest inventories conducted in French Guiana. The dataset comprised 2507 plots (0.4 to 0.5 ha) of undisturbed rainforest distributed over the whole region. After checking the uncertainties of estimates obtained from these data, we used half of the dataset to develop explicit predictive models including spatial and environmental effects and tested the accuracy of the resulting maps according to their resolution using the rest of the data. Forest inventories provided accurate AGB estimates at the plot scale, for a mean of 325 Mg.ha^-1. They revealed high local variability combined with a weak autocorrelation up to distances of no more than10 km. Environmental variables accounted for a minor part of spatial variation. Accuracy of the best model including spatial effects was 90 Mg.ha^-1 at plot scale but coarse graining up to 2-km resolution allowed mapping AGB with accuracy lower than 50 Mg.ha^-1. Whatever the resolution, no agreement was found with available pan-tropical reference maps at all resolutions. We concluded that the combined weak autocorrelation and weak environmental effect limit AGB maps accuracy in rainforest, and that a trade-off has to be found between spatial resolution and effective accuracy until adequate “wall-to-wall” remote sensing signals provide reliable AGB predictions. Waiting for this, using large forest inventories with low sampling rate (<0.5%) may be an efficient way to increase the global coverage of AGB maps with acceptable accuracy at kilometric resolution.