Visual Attention Is Gated at Primary Cortical Afferents
Iris Grothe, David Rotermund, Simon David Neitzel, Sunita Mandon, Udo Alexander Ernst, et al.
(see pages 3441–3452)
We are constantly bombarded by sensory input, but the brain has the power to focus attention on a particular object. Where selection for an attended object occurs remains unknown, but a classic study in visual cortex provides some clues (Moran and Desimone, 1985). Neurons in the V4 area responded to two simultaneously presented stimuli with a firing pattern that was intermediate between the patterns that might be generated by either stimulus alone. But when attention was focused on one stimulus, the V4 firing pattern represented the attended object. Researchers continue to debate whether signals are gated selectively at the level of afferent cortical neurons, or whether selectivity is achieved by modulation of outputs to downstream neurons.
A model quantitatively reproducing experimental observations shows high spectral coherence between attended flickering signals (left) and LFPs recorded in V4 compared with non-attended signals (right).
This week, Grothe, Rotermund, et al. investigated that question by examining local field potentials (LFPs) in V4 as monkeys attended to one of two similar, simultaneously presented stimuli that activated the same set of V4 neurons. The researchers “tagged” the stimuli by independently modulating the luminance of each—that is, the stimuli randomly flickered in brightness. This allowed the researchers to infer which stimulus was driving V4 activity by calculating the spectral coherence between the flicker and the LFP recording.
Analysis of the coherence data indicated that when neither stimulus was attended to, the stimuli contributed equally to V4 LFPs, similar to the classic experiment. But when monkeys attended to one stimulus or the other, the attended signal preferentially drove V4 responses: coherence between LFPs and signals was higher for attended stimuli than for non-attended stimuli. These results suggest that attention modulates synaptic efficacy of specific inputs to V4.
Finally, the researchers created a computational model based on their data from the attended and non-attended neurons in V1 and the local V4 population. The model suggested that attention may be modulated by shifting the phase synchrony of gamma-band oscillations between the V4 population, because V1 neurons representing the attended stimulus oscillated in phase with the V4 neurons, whereas V1 neurons representing the unattended stimulus were in anti-phase with V4 neurons. The study suggests that attention can be gated at early sensory-processing stations.
Distinct Neural Populations Respond to Natural, Drug Rewards
Simone Pfarr, Laura Schaaf, Janine K. Reinert, Elisabeth Paul, Frank Herrmannsdörfer, et al.
(see pages 3507–3519)
Reward-seeking behaviors are influenced by learning from previous experience about contexts and cues that lead to pleasure. That learning depends in part on the medial prefrontal cortex (mPFC), a brain region necessary for self-control and decision-making. The infralimbic area (IL) of the mPFC, in particular, is a key driver of alcohol seeking, as well as seeking of a number of other substances. Unclear, however, is whether IL neuronal ensembles are dedicated to a particular rewarding substance, or whether populations can respond to multiple rewards. This week, Pfarr et al. wanted to investigate responses to natural and drug rewards in the same animal.
Male rats were trained to press two levers: one that delivered 10% ethanol, and another that provided a non-drug reward of a sweet saccharine solution. Orange and lemongrass oils provided contextual olfactory cues during training. Once trained, the rats' motivation was evaluated using a progressive ratio schedule in which animals had to press levers more and more times to receive the same reward, which showed that alcohol was slightly more rewarding than sweetness. Animals underwent 5 d of extinction, and 2 weeks later were presented with olfactory cues to reinstate reward-seeking behaviors.
To identify neuronal populations activated by the rewards, the researchers labeled cFos, a marker of neuronal activity. To distinguish between the neural ensembles activated by seeking drug or saccharine, the researchers used double-cFos mRNA fluorescent in situ hybridization. They stained for two isoforms of cFos—a nascent form detectable 5 min after a behavioral task, and the mature form that appears 30 min after a task. The behaviors recruited distinct neural populations within the IL that were similar in size and organization and overlapped by ∼50%, suggesting they contained cells that respond broadly to reward as well as reward-specific cells. The findings provide new insights to cortical organization of context- and cue-driven memory.
Footnotes
This Week in The Journal was written by Stephani Sutherland, Ph.D.