Does experience provide a permissive or instructive influence on the development of direction selectivity in visual cortex?

Abstract

In principle, the development of sensory receptive fields in cortex could arise from experience-independent mechanisms that have been acquired through evolution, or through an online analysis of the sensory experience of the individual animal. Here we review recent experiments that suggest that the development of direction selectivity in carnivore visual cortex requires experience, but also suggest that the experience of an individual animal cannot greatly influence the parameters of the direction tuning that emerges, including direction angle preference and speed tuning. The direction angle preference that a neuron will acquire can be predicted from small initial biases that are present in the naïve cortex prior to the onset of visual experience. Further, experience with stimuli that move at slow or fast speeds does not alter the speed tuning properties of direction-selective neurons, suggesting that speed tuning preferences are built in. Finally, unpatterned optogenetic activation of the cortex over a period of a few hours is sufficient to produce the rapid emergence of direction selectivity in the naïve ferret cortex, suggesting that information about the direction angle preference that cells will acquire must already be present in the cortical circuit prior to experience. These results are consistent with the idea that experience has a permissive influence on the development of direction selectivity.

Keywords: Area 17; Development; Motion; Striate cortex; Thalamocortical.

Fig. 1

Orientation and direction selectivity in visual cortex. a Left: A bar visual stimulus that is swept back and forth across the receptive field of a cortical neuron. The orientation of the bar is varied over 4 angles, and the direction of motion of the bar is varied over 8 directions. Right: Responses to stimulation at different orientations and directions. This particular neuron responds to many orientations and directions, but provides particularly strong responses for stimuli moving up and to the right. b A tuning curve graph of the responses of the same cell as a function of direction angle. The preferred direction and the “null” direction (direction opposite the preferred) are indicated. Adapted from [36, 78]

Fig. 2

In ferrets and primates, direction selectivity develops postnatally. In ferrets, it has been shown that the development of direction selectivity requires visual experience. a Profile of development of orientation selectivity and direction selectivity in ferrets [19] and rough equivalent in macaque from [29]. b The influence of experience on the development of direction selectivity in ferret. Light-reared animals (that is, typically-reared animals) exhibit strong direction selectivity for animals P63 or older [19]. By contrast, dark-reared animals P63 or older exhibit poor selectivity for direction selectivity. Animals that were dark-reared until P45–50 and then reared under typical conditions (“Early dark-reared”) also failed to develop direction selectivity, indicating that early visual experience is required for the proper development of direction selectivity [19]. Animals that were dark-reared only until P35, and then allowed 2–3 weeks of visual experience, exhibited strong direction selectivity [19]. Finally, artificial experience with moving stimuli for 3–6 h is sufficient to cause a rapid increase in direction selectivity in visually naïve ferrets [35]. Adapted from [19]

Fig. 3

Initial biases in the naïve cortex correlate with the direction angle preference that is acquired. a Left: Sketch of imaging field in ferret visual cortex at the onset of visual experience. Neurons exhibit very weak direction selectivity, as indicated by small arrows. Nevertheless, there are regions that have statistically significant biases for particular directions, such as right (green) and left (blue) as shown [35]. These biases are found even in animals that have been dark-reared [23], suggesting that they are formed independent of any visual experience, including that through the closed lids. Middle: Artificial experience of 3–6 h with moving visual stimuli is sufficient to produce the rapid emergence of direction selectivity in visual cortex. In this case, stimuli moved in one of two opposite directions (random alternation), 5 s on, 5 s off, in 20 min blocks, with a 10 min rest period. Right: Sketch of imaging field after bidirectional experience, with enhancement of direction selectivity in both regions [35]. b Left: Sketch of initial imaging field at the onset of visual experience. Middle: Animal is provided with 3–6 h of artificial experience with moving stimuli, but here the stimuli move only in a single direction. Right: Sketch of imaging field after unidirectional experience. Neurons in regions that were biased toward the “trained” direction exhibit robust increases in direction selectivity. Neurons in regions that were biased to the opposite direction showed little change. Neurons in intermediate regions could be recruited to exhibit selectivity for the trained direction [23]

Fig. 4

A feed-forward model with Hebbian plasticity and increasing feed-forward inhibition can, in principle, develop direction selectivity in an instructive manner. a Schematic of the feed-forward model [37]. An array of LGN neurons with a broad array of different position preferences and response latencies provide input to a cortical excitatory neuron and a feed-forward cortical inhibitory neuron. The cortical inhibitory neuron provides input to the excitatory neuron. The picture shows an immature network that has a slight (subthreshold) bias for downward motion (darker LGN cells indicates slightly stronger weight). Connections from LGN to the cortical excitatory neuron undergo spike-timing-dependent plasticity, while the synapse from the cortical inhibitory neuron onto the cortical excitatory neuron increases with each bout of stimulation, forcing competition among the inputs [38]. b Responses from the naïve model cortical neuron and LGN neurons to upward and downward stimuli. Each row of LGN cells responds to stimuli at a particular position, with varying latencies. Each black square represents the spiking activity of a single LGN cell. In the middle of the stimulus, the 5 LGN cells along each diagonal are activated simultaneously, allowing the cortical excitatory neuron to fire. c Connections after hundreds of bidirectional stimulation events. The increasing inhibition has forced the cortical excitatory neuron to develop selectivity for downward motion; the LGN inputs that support upward motion (the direction opposite the initial bias) are eventually weakened, because they do not drive the cell after the feed-forward inhibition has developed to full strength. d After training, the cortical neuron responds to stimulation in the downward direction exclusively. Stimulation with downward motion at the appropriate speed (∆d/∆t) will cause the 5 cells that comprise the diagonal to be activated simultaneously, providing strong drive to the cortical neuron. Stimulation with upward motion activates the same 5 LGN cells, but asynchronously, such that they do not drive the cortical cell. Adapted from [37, 39]

Fig. 5

Hypotheses about initial circuit and the effect of speed training. a Hypotheses about initial circuit. In Possible Juvenile State I, the cortical neuron can receive input from an array of LGN cells with a wide variety of position preferences and latencies. In Possible Juvenile State II, the cortical neuron is pre-constrained to receive inputs from cells with particular position and latency values. b Hypothesized adult state. Neurons with particular position and latency preferences converge on the cortical neuron, resulting in direction selectivity. c Impact of providing experience with stimuli moving at different speeds. In Juvenile State I, only LGN neurons with positions and delays that were activated by a particular speed are strengthened, resulting in speed selectivity that matches the experienced speed. d Speed selectivity before and after training under Juvenile State I. e In Juvenile State II, the eventual speed tuning is built-in before the onset of experience, and visual experience with moving gratings merely enhances this pre-constrained tuning. f Speed selectivity before and after training under Juvenile State II. Experiments in ferrets strongly resembled the outcomes in (e) and (f) [39]. Adapted from [39]

Fig. 6

Direct cortical activation in visually naïve ferrets produced an increase in direction selectivity. a Engineered viruses that cause the expression of Channelrhodopsin-2 were injected in young ferrets several days before eye opening. b Initial orientation and direction selectivity were assessed with visual stimuli. c Next, the visual display was switched off and patterned light was shone on the cortex to produce specific activity patterns for several hours. d Periodically, the visual display was switched on and orientation and direction selectivity were assessed with visual stimulation. We found that 9 h of this direct cortical stimulation protocol resulted in an increase in cortical direction selectivity [40]. Adapted from [40]