The hard and soft filter models: what do they say about attention?

People are constantly subjected to very complex situations in which a large number of stimuli compete for our attention. Although we don't realize it, we spend a lot of time sorting out the relevant from the irrelevant, separating the wheat from the chaff.

This is mainly due to the fact that our resources to process information are very limited, so if we were to open the dam of our attention without any control we would end up feeling how the ability to understand what is happening overflows around.

In order to find out how our brain works in situations as frequent as this, they were postulated throughout the s. XX a series of hypotheses that would mark the way forward over the years. Of this, the rigid and attenuated filter model was the pioneer.

In this article we will address and specify the postulates of this classic model, placing special emphasis on the different points through which information passes from when it is perceived by the senses until it is stored persistently in the memory.

• Related article: "History of Psychology: authors and main theories"

Hard Filter Model and Soft Filter Model

The rigid filter model and the attenuated filter model propose a dynamic for the functioning of attention that stands out for inserting a filter or screening mechanism, through which the complexity of the environment would be refined and what was relevant would be selected from it. It includes elements of the multistore theory on memory, whose previous knowledge is basic for the correct understanding of these models: sensory storage, short-term memory and long-term memory term.

1. sensory warehouse

The sensory warehouse is the first stop in information processing, as it is the space in which sensations from the sense organs are deposited.

The perceptive fact, through any of its different modalities (visual, acoustic, olfactory, gustatory and tactile), requires little time to be captured by the nervous system, but requires a somewhat more elaborate analysis to determine its physical properties and nuances.

In this warehouse, with a very large capacity but a very limited duration, an extraordinary volume of items settles on the situation in which we find ourselves, although almost all of them dissolve in a few seconds (without mediating a cognitive analysis deep). The information would be transferred from here to short-term memory, after being sifted by the attentional filter, which will be discussed in detail later.

2. short term memory

After the information coming from the senses has crossed the aforementioned sensory store, it would be projected into short-term memory. At the moment an abstraction of the sensory image is retained, a kind of interpretation of the object on which the attention was placed.

This interpretation is an inaccurate picture, since has been subjected to a first process of cognitive elaboration in which some of its objective properties may have been altered.

This memory has a smaller amplitude than the sensory store, but its duration is much longer. In this way, the (now conscious) retention of this data can be prolonged for a few minutes, but it will tend to dissolve if it is valued as irrelevant by the receiver. In general terms, it is estimated that an individual (under normal circumstances) can retain up to seven single elements at this processing station, the normal range being three to eleven.

The anterograde amnesia provides reliable information about the very existence of this store, and is one of the arguments most commonly used by defenders of the compartmentalization of memory. This phenomenon describes the formation of new learning that only lasts a few minutes, after which they disappear without being consolidated in any case (so they would never enter long-term storage).

• You may be interested in: "Types of memory: how does the human brain store memories?"

3. long term memory

When the information has been perceived by the sense organs, sent to the sensory store and derived to short-term memory term, there is a process of conscious analysis of its importance in order to transfer it to the last station: long-term memory. term. It is in this place where the declarative memories that are far away in time live, and to which we turn voluntarily when we wish.

Long-term memory has an indefinite duration and can last a lifetime. Here is stored a declarative crystallization of the events experienced (episodic), the knowledge about the world (semantic) and the skills acquired (procedural); all of this is necessary due to its emotional relevance and/or its adaptive value. There are many brain regions involved in itTherefore, it is usually affected during the evolution of dementia processes.

• You may be interested in: "Parts of the human brain (and functions)"

filter models

Once the different stores into which the memory is divided are known, and after the analysis of its process since the object is captured through the senses until it is eventually stored in a durable form, it is easier to understand the rigid filter model and attenuated. These theories were developed in order to understand the way in which a human being deals with complex situations in which very diverse information competes with each other to be perceived, processed and stored.

Thus, it explores the characteristics of selective attention: how we discriminate information from the environment when this is complex, in order to collect what is relevant and articulate suitable responses according to the context. Here we will review two pioneering hypotheses on this matter: the hard (Donald Broadbent) and soft (Anne Treisman) filter, both being the theoretical foundation on which subsequent theoretical elaborations would be built (such as the late filter model or others).

To get closer to these models, the most useful thing is to give an example: let's imagine that we are meeting with a friend in a bar, drinking coffee, while they tell us an interesting story. How do we focus attention on his words if the environment is flooded with other sounds that compete with them (such as people talking, silverware clinking, and even cars driving near where we are)?

In order to explore what happens in our brain in everyday situations like this, the authors used an experimental type procedure known as dichotic listening, and which consists of the simultaneous emission of two different messages through each of the auditory channels (with the help of headphones). The participant would remain seated listening to its content (numbers, words, etc.), and after the presentation would point out what he thinks he has perceived.

With this simple method, the dynamics of selective attention could be explored., one of the expressions of this executive function, which consists of choosing a relevant stimulus and omitting irrelevant ones when both are presented at the same time. It is a basic skill for the development of activities of daily living, together with attention sustained (or vigilance) and divided (efficient approach to two or more important tasks at the same time). time).

While it is true that both Broadbent and Treisman agreed on basic aspects, such as the existence of a sensory store and the process of transmission of information from short-term memory to long-term storage, showed some discrepancies related to the concept of "filter". In both cases their existence was considered as a phase of prior screening of the complexity stimulate, but different views related to its degree of permeability were maintained (as will be seen later).

1. rigid filter model

The use of a filter could be likened, in Broadbent's own words, to the "neck of a bottle". Although the stimulus field in which we find ourselves may be very complex, our cognitive abilities only allow a discrete percentage of this to be processed and analyzed without exceeding the resources of which we have. For this purpose, the filter would act as a sieve for environmental diversity to translate it into clear, operational, and manageable terms.

This filter would be located, according to the author (although it was later questioned from the framework of the late Deutsch and Deutsch filter), right at the end of sensory storage and before short-term memory. In this way, the stimuli would be processed in series, and never in parallel (which implies that the information is analyzed one by one and never simultaneously). With this filter, a selection of the relevant and the irrelevant would be facilitated, so that the former would transcend into short-term memory and the latter would be radically omitted.

According to Broadbent, the screening criterion would be the physical property of the stimulus, such as the tone or volume of the human voice, as well as the unpredictability with which it broke into the perceptual field. Be that as it may, from these variables the individual would choose what is relevant to him, while the rest of the elements would be completely ignored without being attended to or understood.

Broadbent provided empirical evidence through dichotic listening, through an experimental condition that consisted of the emission of a short list of numbers in each of the evaluand's ears. For example, if you heard the sequence 947 through your left ear and 246 through your right ear, you would only remember one or other (but never information that combined the two sources or all the items included in the trial). He concluded that each of the ears would function as an independent channel, with only one being selected and the other omitted entirely.

2. Attenuated filter model

The smooth filter was proposed by Treisman, following his attempts to replicate Broadbent's findings. There is a basic difference between the proposals of these two authors, located precisely in the qualities of the filter as an inserted element within the processing of information.

Treisman considered that there was no absolute block of the unattended stimulus., but that this was processed in some way despite the fact that the person tried to focus on what was relevant. Unattended messages would see his salience lowered, but would not disappear.

Like Broadbent, he used dichotic listening to test his hypothesis. In this case, verbal messages (phrases with meaning) were used, but dividing the informative segments in a particular way.

For example, through the left ear two messages without logical connection would be reproduced successively (such as "I took a coat we caught four fish”), while on the right side another one would sound very similar in terms of structure (“we went fishing because it was cold"). In such a case, the person would say hearing "I took a coat because it was cold" or "we went fishing and caught four fish", showing that he had attended to both messages at the same time.

The explanation for this finding for Treisman was that the filter does not completely nullify the unattended message, but it continues to be processed at some level and can come to attract attention if it brings consistency to what was being perceived up to that moment. It also showed, for example, that people remembered basic aspects of "ignored" information, even using Broadbent's own paradigm (changes in the volume of the voice, timbre, tone or gender of the announcer; as well as the reproduction of the name of the evaluated subject).

Thus, certain conditions of the individual (such as their vital experience or their expectations of the future) would be responsible for attributing perceptual relevance to the stimulus. Furthermore, the filter would act by weakening less relevant messages, but these would not be completely inhibited (as suggested by the hard filter). Therefore, there would be a basic processing at the semantic level (of a pre-categorical type) with which the selection tasks would be optimized without saturating the cognitive system.

Bibliographic references:

• Driver, J. (2001). A selective review of selective attention research from the past century. British Journal of Psychology, 92, 53-78.
• Lee, K. and Choo, H. (2011). A critical review of selective attention: An interdisciplinary perspective. Artificial Intelligence Review, 40(1), 27-50.