Home
This Title All WIREs
WIREs RSS Feed
How to cite this WIREs title:
WIREs Cogn Sci
Impact Factor: 3.476

Neural and BOLD responses across the brain

Full article on Wiley Online Library:   HTML PDF

Can't access this content? Tell your librarian.

Abstract Functional Magnetic Resonance Imaging (fMRI) has quickly grown into one of the most important tools for studying brain function, especially in humans. Despite its prevalence, we still do not have a clear picture of what exactly the blood oxygenation level dependent (BOLD) signal represents or how it compares to the signals obtained with other methods (e.g., electrophysiology). We particularly refer to single neuron recordings and electroencephalography when we mention ‘electrophysiological methods’, given that these methods have been used for more than 50 years, and have formed the basis of much of our current understanding of brain function. Brain function involves the coordinated activity of many different areas and many different cell types that can participate in an enormous variety of processes (neural firing, inhibitory and excitatory synaptic activity, neuromodulation, oscillatory activity, etc.). Of these cells and processes, only a subset is sampled with electrophysiological techniques, and their contribution to the recorded signals is not exactly known. Functional imaging signals are driven by the metabolic needs of the active cells, and are most likely also biased toward certain cell types and certain neural processes, although we know even less about which processes actually drive the hemodynamic response. This article discusses the current status on the interpretation of the BOLD signal and how it relates to neural activity measured with electrophysiological techniques. WIREs Cogn Sci 2012, 3:75–86. doi: 10.1002/wcs.153 This article is categorized under: Neuroscience > Physiology

High resolution gradient echo (GE)‐BOLD (a), spin‐echo (SE)‐BOLD (b), functional cerebral blood flow (CBF) (c), and functional cerebral blood volume (SE‐CBV) (d) in V1 of anesthetized monkeys. The stimulus was a black‐and‐white rotating checkerboard. The maximum percentage change for GE‐BOLD is located at the cortical surface while the maximum SE‐BOLD, CBV, and CBF signal occurs within the cortex at the level of cortical layer IV. Data were acquired at 4.7 T (GE‐BOLD, CBF, CBV) and 7 T (SE‐BOLD). In‐plane resolution was 333 × 333 µm2 for GE‐, SE‐BOLD, and CBV and 375 × 333 µm2 for CBF. Functional CBF was measured using continuous arterial spin labeling (CASL) and functional CBV was measured using SE‐echo planar imaging (EPI) after injection of Monocrystalline iron oxide nanocolloid (MION). See Goense et al.16,17 and Zappe et al.18 for experimental details. (Reprinted with permission from Ref 19. Copyright 2008 Nature Publishing Group; Ref 16. Copyright 2010 Elsevier Ltd.; Ref 18. Copyright 2008 Nature Publishing Group)

[ Normal View | Magnified View ]

Dissociation of the multiunit activity (MUA) and BOLD response in V1 of an awake monkey. (a) Functional activation maps in response to a 6° rotating checkerboard stimulus (inset in b). The arrow marks the location of the electrode. (b) The average MUA, local field potentials (LFP) (20–60 Hz) and BOLD time courses show that while the LFP stayed elevated for the duration of the stimulus, the MUA rapidly returned to baseline after a transient onset response. The prolonged time course of the BOLD response suggests a more sustained driving mechanism of the BOLD response as opposed to the transient MUA signal. The dotted line indicates the regressor, i.e. the neural signal convolved with the theoretical hemodynamic response function (HRF), which shows that the MUA‐derived regressor cannot capture the sustained part of the BOLD response. (Reprinted with permission from Ref 2. Copyright 2008 Elsevier Ltd.)

[ Normal View | Magnified View ]

(a) Estimating the impulse response function (IRF) between neural [electroencephalography (EEG)] and hemodynamic (fMRI) activity in human primary visual cortex. (b) Using inverse modeling, the EEG source activity is estimated within the primary visual cortex (yellow square), and compared to the corresponding BOLD signal from the same subject. For both signals, amplitude is given in standard deviation (SD) units. The temporal resolution of the EEG source waveform is high (200 Hz), while that of the BOLD is low (0.4 Hz). (c) The IRF, which reveals the temporal structure between the two signals, peaks at approximately 6 s and returns to baseline at about 15 s. This agrees very well with previous local field potentials (LFP)‐BOLD measurements in macaque visual cortex.1 Computing the IRF between EEG source activity and BOLD may be a useful way to study neurovascular coupling across different brain areas. (Reprinted with permission from Ref 97. Copyright 2010 Elsevier Ltd.)

[ Normal View | Magnified View ]

Browse by Topic

Neuroscience > Physiology

Access to this WIREs title is by subscription only.

Recommend to Your
Librarian Now!

The latest WIREs articles in your inbox

Sign Up for Article Alerts