Thinking of memory as a network is obvious for two reasons. First, it’s the image we all hold in our minds of the brain as a ball of neurons networked together. Second, the Internet has become a popular model for the brain, particularly as social networking has begun to transform the business of marketing.

As early as the sixties, artificial intelligence researchers used semantic networks as a way of representing artificial memory. Neural networks, widely used in analyzing huge databases and machine learning, are an extension of this early work.

Semantic maps are usually created to display the relationships between words or concepts. A fresh approach is to show the relationship between visual images. The example below is taken from the data in a pretest of a television commercial. Shortly after an audience has watched the ad and processed it into memory, it quantifies the complex, non-linear interrelationships that exist between the different images in a TV commercial.

To build such a memory network is a straightforward process:
First, the different elements from an ad (it can be copy, not just spoken words) are retrieved from a sample of target consumer memories, using self-report research methods that are easily done over the Internet.

Second, each element from the ad is “tagged” with a number representing a metric obtained in the research, such as the strength of memory associated with each element, the positive or negative feelings associated with each element, the conceptual, strategic meaning associated with each element or the degree of brand affinity of each element.

Third, a correlation analysis between the various numerical tags is performed. The results are displayed as a network model of the associative strength of the different advertising elements obtained from consumer memory.

A network is a useful model of advertising-created memories because it underscores the important fact that remembering is a complex process of consolidating our perceptions. A commercial is a linear sequence of images presented in time and frequently organized with narrative structure–but it is folded and translated into a brand memory that is widely distributed over the networks of the brain.

On a larger scale, the network model of memory might even be taken one step further. It is interesting to speculate about the implications of the billions of personal photographs that are being uploaded every day to social networks like Facebook. They now represent a form of collective memory for our society and our culture.