Limits...
Perceived intensity of somatosensory cortical electrical stimulation.

Fridman GY, Blair HT, Blaisdell AP, Judy JW - Exp Brain Res (2010)

Bottom Line: Artificial sensations can be produced by direct brain stimulation of sensory areas through implanted microelectrodes, but the perceptual psychophysics of such artificial sensations are not well understood.We then conducted a series of two-alternative forced choice behavioral experiments in which we systematically tested the ability of rats to discriminate frequency, amplitude, and duration of electrical pulse trains delivered to the whisker barrel somatosensory cortex.We found that the model was able to predict the performance of the animals, supporting the notion that perceived intensity can be largely accounted for by spatiotemporal integration of the action potentials evoked by the stimulus train.

View Article: PubMed Central - PubMed

Affiliation: Biomedical Engineering Department, UCLA, Los Angeles, CA 90095, USA. gfridma1@jhmi.edu

ABSTRACT
Artificial sensations can be produced by direct brain stimulation of sensory areas through implanted microelectrodes, but the perceptual psychophysics of such artificial sensations are not well understood. Based on prior work in cortical stimulation, we hypothesized that perceived intensity of electrical stimulation may be explained by the population response of the neurons affected by the stimulus train. To explore this hypothesis, we modeled perceived intensity of a stimulation pulse train with a leaky neural integrator. We then conducted a series of two-alternative forced choice behavioral experiments in which we systematically tested the ability of rats to discriminate frequency, amplitude, and duration of electrical pulse trains delivered to the whisker barrel somatosensory cortex. We found that the model was able to predict the performance of the animals, supporting the notion that perceived intensity can be largely accounted for by spatiotemporal integration of the action potentials evoked by the stimulus train.

Show MeSH

Related in: MedlinePlus

Perturbing SI model alters perceptual model predictions. A We used three different values for the time constant τ in the SI model for animal 1. We fit the three models which varied by the value of the time constant, to the frequency testing experiment animal performance (same experimental data as in Fig. 4) on the left to obtain the behavioral model parameters a,b,c, and d for each model variation (the model fits overlay each other). We then used the three models which now had the complete set of parameters to predict animal performance on the duration and amplitude testing experiments (same experimental data as in Fig. 5). The three model predictions are shown on the duration and on the amplitude testing performance plots on the right. Varying SI time constant parameter from 0.48 s, obtained through optimization, caused changes to the ability of the model to predict the animal behavior on the duration task and to a lesser degree on the amplitude task (the model prediction for τ = 1.1 s overlays the model prediction for τ = 0.48 s). Analogously, in B, when we varied the SI exponent from the theoretically derived value of 3/2, we observed a deviation from the more optimal prediction of animal performance on the amplitude experiment; with no changes to the predictions on the duration experiment performance (the model predictions using all three SI exponents overlay each other)
© Copyright Policy
Related In: Results  -  Collection


getmorefigures.php?uid=PMC2875471&req=5

Fig6: Perturbing SI model alters perceptual model predictions. A We used three different values for the time constant τ in the SI model for animal 1. We fit the three models which varied by the value of the time constant, to the frequency testing experiment animal performance (same experimental data as in Fig. 4) on the left to obtain the behavioral model parameters a,b,c, and d for each model variation (the model fits overlay each other). We then used the three models which now had the complete set of parameters to predict animal performance on the duration and amplitude testing experiments (same experimental data as in Fig. 5). The three model predictions are shown on the duration and on the amplitude testing performance plots on the right. Varying SI time constant parameter from 0.48 s, obtained through optimization, caused changes to the ability of the model to predict the animal behavior on the duration task and to a lesser degree on the amplitude task (the model prediction for τ = 1.1 s overlays the model prediction for τ = 0.48 s). Analogously, in B, when we varied the SI exponent from the theoretically derived value of 3/2, we observed a deviation from the more optimal prediction of animal performance on the amplitude experiment; with no changes to the predictions on the duration experiment performance (the model predictions using all three SI exponents overlay each other)

Mentions: We compared the predictive ability of three SI models, which varied only by values of τ. We used 1.1, 0.1 s, and the optimally derived value of 0.48 s for τ. For each model, we found the best fit behavioral function such that we obtained least-squares fit to the frequency testing experiment (Fig. 6A, left plot). The model fitting optimized the functions well enough to overlay them on top of each other on the plot. With the three sets of behavioral function parameters corresponding to each SI model, we then used the three complete models to predict behavior on the duration and amplitude testing experiments (Fig. 6A, right plots). We observed that the changes in the time constant caused substantial changes to the ability to predict behavior on the duration testing experiment and to a lesser extent on the amplitude testing experiment.Fig. 6


Perceived intensity of somatosensory cortical electrical stimulation.

Fridman GY, Blair HT, Blaisdell AP, Judy JW - Exp Brain Res (2010)

Perturbing SI model alters perceptual model predictions. A We used three different values for the time constant τ in the SI model for animal 1. We fit the three models which varied by the value of the time constant, to the frequency testing experiment animal performance (same experimental data as in Fig. 4) on the left to obtain the behavioral model parameters a,b,c, and d for each model variation (the model fits overlay each other). We then used the three models which now had the complete set of parameters to predict animal performance on the duration and amplitude testing experiments (same experimental data as in Fig. 5). The three model predictions are shown on the duration and on the amplitude testing performance plots on the right. Varying SI time constant parameter from 0.48 s, obtained through optimization, caused changes to the ability of the model to predict the animal behavior on the duration task and to a lesser degree on the amplitude task (the model prediction for τ = 1.1 s overlays the model prediction for τ = 0.48 s). Analogously, in B, when we varied the SI exponent from the theoretically derived value of 3/2, we observed a deviation from the more optimal prediction of animal performance on the amplitude experiment; with no changes to the predictions on the duration experiment performance (the model predictions using all three SI exponents overlay each other)
© Copyright Policy
Related In: Results  -  Collection

Show All Figures
getmorefigures.php?uid=PMC2875471&req=5

Fig6: Perturbing SI model alters perceptual model predictions. A We used three different values for the time constant τ in the SI model for animal 1. We fit the three models which varied by the value of the time constant, to the frequency testing experiment animal performance (same experimental data as in Fig. 4) on the left to obtain the behavioral model parameters a,b,c, and d for each model variation (the model fits overlay each other). We then used the three models which now had the complete set of parameters to predict animal performance on the duration and amplitude testing experiments (same experimental data as in Fig. 5). The three model predictions are shown on the duration and on the amplitude testing performance plots on the right. Varying SI time constant parameter from 0.48 s, obtained through optimization, caused changes to the ability of the model to predict the animal behavior on the duration task and to a lesser degree on the amplitude task (the model prediction for τ = 1.1 s overlays the model prediction for τ = 0.48 s). Analogously, in B, when we varied the SI exponent from the theoretically derived value of 3/2, we observed a deviation from the more optimal prediction of animal performance on the amplitude experiment; with no changes to the predictions on the duration experiment performance (the model predictions using all three SI exponents overlay each other)
Mentions: We compared the predictive ability of three SI models, which varied only by values of τ. We used 1.1, 0.1 s, and the optimally derived value of 0.48 s for τ. For each model, we found the best fit behavioral function such that we obtained least-squares fit to the frequency testing experiment (Fig. 6A, left plot). The model fitting optimized the functions well enough to overlay them on top of each other on the plot. With the three sets of behavioral function parameters corresponding to each SI model, we then used the three complete models to predict behavior on the duration and amplitude testing experiments (Fig. 6A, right plots). We observed that the changes in the time constant caused substantial changes to the ability to predict behavior on the duration testing experiment and to a lesser extent on the amplitude testing experiment.Fig. 6

Bottom Line: Artificial sensations can be produced by direct brain stimulation of sensory areas through implanted microelectrodes, but the perceptual psychophysics of such artificial sensations are not well understood.We then conducted a series of two-alternative forced choice behavioral experiments in which we systematically tested the ability of rats to discriminate frequency, amplitude, and duration of electrical pulse trains delivered to the whisker barrel somatosensory cortex.We found that the model was able to predict the performance of the animals, supporting the notion that perceived intensity can be largely accounted for by spatiotemporal integration of the action potentials evoked by the stimulus train.

View Article: PubMed Central - PubMed

Affiliation: Biomedical Engineering Department, UCLA, Los Angeles, CA 90095, USA. gfridma1@jhmi.edu

ABSTRACT
Artificial sensations can be produced by direct brain stimulation of sensory areas through implanted microelectrodes, but the perceptual psychophysics of such artificial sensations are not well understood. Based on prior work in cortical stimulation, we hypothesized that perceived intensity of electrical stimulation may be explained by the population response of the neurons affected by the stimulus train. To explore this hypothesis, we modeled perceived intensity of a stimulation pulse train with a leaky neural integrator. We then conducted a series of two-alternative forced choice behavioral experiments in which we systematically tested the ability of rats to discriminate frequency, amplitude, and duration of electrical pulse trains delivered to the whisker barrel somatosensory cortex. We found that the model was able to predict the performance of the animals, supporting the notion that perceived intensity can be largely accounted for by spatiotemporal integration of the action potentials evoked by the stimulus train.

Show MeSH
Related in: MedlinePlus