Neuromarketing Prediction via Neural Correlation Mapping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing neuromarketing methods are ineffective in predicting diverse behavioral responses of a large audience, as they rely on single measures of correlation and do not consider combining neural signals with additional information such as stimulus properties or behavioral responses from groups of individuals.
Innovation Solution
A method involving the use of intra- and inter-subject correlations in neurological data to predict behavioral responses, which includes receiving behavioral and neurological data from populations, processing this data to identify commonality, and applying a mapping to predict responses to new stimuli, utilizing techniques like correlated components analysis for dimensionality reduction and machine learning for accurate predictions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If single brain region activity is used as a marker of consumption, then the measurement is simple, but the prediction accuracy is insufficient for complex tasks
Solution Approach 1:
The patent segments the brain into multiple distinct regions (e.g., OFC, vmPFC, insula, amygdala, hippocampus) and measures activity in each region separately. This segmentation allows the system to capture the distributed neural networks involved in complex consumption tasks while maintaining the simplicity of region-specific measurements. Each region's activity serves as a specialized marker for different aspects of the consumption experience.
Solution Approach 2:
The patent combines activity measurements from multiple brain regions into an integrated predictive model. By merging the signals from different neural regions that are known to be involved in consumption processing, the system achieves higher prediction accuracy for complex tasks while still relying on relatively simple regional activity measurements as input components.
2Loss of information
If arbitrary mapping from high-dimensional neural data to behavior is attempted, then comprehensive neural information is captured, but the mapping fails due to high noise and limited data
Solution Approach 1:
The patent transforms the high-dimensional neural data by changing the parameters from raw activity levels to correlation coefficients. Specifically, it computes the correlation between each subject's neural activity pattern and the group average pattern, producing a standardized metric that ranges from -1 to 1. This parameter transformation reduces noise and enables reliable mapping even with limited data by focusing on the consistent, stimulus-related component of neural activity.
Solution Approach 2:
The patent extracts the stimulus-specific component of neural activity by computing correlation with the group average. This extraction process separates the meaningful signal (stimulus-driven response) from the noise (individual variability and ongoing activity not specific to the stimulus). The resulting correlation measure represents only the reliable, stimulus-related neural response, enabling accurate behavioral prediction.
3Use of energy by moving object
If stimulus-decoupled ongoing activity is included in neural measurements, then signal amplitude is strong, but the information relevant to neuromarketing is diluted
Solution Approach 1:
The patent extracts only the stimulus-coupled component of neural activity by computing the correlation between individual subject activity and the group average activity. This extraction automatically separates the stimulus-specific signal from the stimulus-decoupled ongoing activity. The resulting correlation measure contains only the information relevant to the neuromarketing question (how the stimulus affected the brain), while filtering out the irrelevant ongoing activity that would otherwise dilute the signal.
Data Source
AI summary
A method of predicting response to a sensory stimulus includes, with a processor, automatically receiving behavioral data representing the response of a first population of subjects to a reference stimulus. Data representing the neurological responses of a second, different population of subjects to the reference sensory stimulus are received and processed to provide group-representative data indicating commonality between the neurological responses of at least two members of the second population. A mapping from the group-representative data to the received behavioral data is produced. Test data representing the neurological responses of a third population of subjects to a test sensory stimulus are received and processed to provide test group-representative data indicating commonality between the neurological responses to the test sensory stimulus of at least two members of the third population. The mapping is applied to the test group-representative data to provide predicted behavioral data.