[escepticos] Re-interpretando señales de fMRI

[Carlos Dominguez]

Saludetes
Carolus

It is assumed that the changes in blood flow measured during functional
brain-imaging studies reflect changes in neural activity. Two
recent studies that directly measured blood flow, blood oxygenation and
neural activity in animals suggest that this assumption must be reconsidered
at least* in some circumstances.*

Devor et al. measured blood oxygenation (using spectral imaging of
oxyhaemoglobin, deoxyhaemoglobin and total haemoglobin)
and blood flow (using laser speckle imaging) in the somatosensory cortex of
anaesthetized rats in response to forepaw stimulation.
As expected, both measures increased in the contralateral cortex but
decreased on  the ipsilateral side. Surprisingly, however, the
uptake of labelled 2-deoxyglucose, a measure of metabolic activity, in
response to forepaw stimulation increased in both the contra- and the
ipsilateral cortex. Moreover, electrophysiological recordings revealed a
stimulus-evoked increase in spiking in both sides of the somatosensory
cortex. Thus, in the ipsilateral side, blood flow and oxygenation were
decoupled from glucose consumption and neuronal activity, suggesting that
energy consumption associated with neuronal activity does not necessarily
determine blood flow.
Sirotin and Das developed an optical imaging technique that allowed them to
measure both blood volume and oxygenation in alert monkeys while
simultaneously recording neural activity. Their experiments took place in a
dark room, with a tiny point
of light serving as a visual cue. The colour of this light point alternated
at regular intervals; the colours indicated either a ‘fixate’ phase, in
which the monkeys had to stare at the light point for several seconds in
order to receive a juice reward, or a ‘relax’ phase.

Importantly, the light point was visible at all times and the cortical area
receiving input from the light point fell outside of the area that was being
imaged; thus, any observed haemodynamic or neural-activity changes were
assumed to be trial-related rather than visually driven.

Blood volume and oxygenation in the primary visual cortex (V1) showed robust
periodic changes that followed the rhythm of the trial phases. There was no
temporal correlation between these changes and multi-unit spiking and local
field-potential measurements in the V1, whereas the haemodynamic signals
correlated closely with these measurements in trials in which strong visual
stimuli, such as coloured gratings, were used in addition to the light
point. Thus, stimulus-induced haemodynamic signals reflected
neuronal activity, but the fluctuating, trial related haemodynamic signals
observed in this study did not.

What then underlied the fluctuating haemodynamic signals? The authors found
that the signals were directly linked to the
timing of the task trials: changing the duration of the trials resulted in
an equivalent change in the timing of the haemodynamic
signal. Moreover, the signal changes began before the onset of a trial,
indicating that they were not triggered by the colour change of the light
point; rather, they might have reflected anticipatory responses.

*These studies will urge a rethink of the interpretation of functional MRI
signals as they show that an increased haemodynamic
signal does not always reflect increased neural activity*. Sirotin and Das
showed that, *depending on the type of stimulus used*, *the signal might be
anticipatory rather than responsive*, whereas Devor et al. showed that
neural activity might in some cases result in vasoconstriction rather than
dilation, and thus in reduced blood flow.

Leonie Welberg

ORIGINAL RESEARCH PAPERS Devor, A. et al. Stimulus-induced changes in blood
flow and 2-deoxyglucose uptake
dissociate in ipsilateral somatosensory cortex. J. Neurosci. 28, 14347–14357
(2008) | Sirotin, Y. B. & Das, A. Anticipatory
haemodynamic signals in sensory cortex not predicted bylocal neuronal
activity. Nature 457, 475–479 (2009)