"Attention Training" via Meditation Influences the Ventral and Dorsal Attentional Networks Differently

As discussed earlier this week, meditation may be an alternative form of brain training - or "brain untraining" - that shows transfer to tasks requiring cognitive control. There have been a few updates to this fascinating line of research, not least of which is a fascinating paper by Amishi Jha and colleagues from the University of Pennsylvania. They showed that relative to a control group, meditation influences particular components of attention in ways that are compatible with beliefs long held in the meditation community.

In particular, Jha et al focus on mindfulness meditation, which is defined as continuous and "non-judgemental" attention to the unfolding present, and has shown a variety of beneficial health effects (as described in a previous post). The meditation literature suggests there are two forms of mindfulness meditation: a "concentrative" form which involves continuous allocation of attention (e.g., towards breathing) and a "receptive" form which involves keeping attention in a state of readiness of preparedness. Interestingly, many meditation protocols suggest that the concentrative form must be developed first, due to the benefits it provides in limiting "mind wandering" during receptive meditation.

Jha et al connect these rather opaque terms to cognitive neuroscience with recourse to the Corbetta & Shulman model of attention. Jha et al suggest that "Receptive meditation" might involve the ventral attentional network (involved in monitoring the environment) whereas the "concentrative meditation" might involve the dorsal attentional network (involved in directing attention to particular stimuli or responses).

To test this idea, Jha et al recruited three groups of subjects: a control group of nursing students from UPenn, a "training" group of meditation-naive subjects who would undergo 8 weekly 3-hour sessions of meditation training, and a "retreat" group of experienced meditators who would go on a month-long mindfulness meditation retreat.

Jha et al gave all subjects both pre- and post-tests on Posner and colleagues' ANT task, described in detail in italics (casual readers may wish to skip this).

The ANT dissociates the ventral system from two aspects of the dorsal system. Roughly speaking, the ventral system is assessed by an "alerting" score on the ANT, which measures the degree of impairment subjects suffer when they are not warned that a target stimulus is upcoming (thus assessing their ability to maintain "receptive" or "alert" attention, perhaps by monitoring the environment). Similarly, the dorsal network is assessed at both the stimulus and response levels. The stimulus-level "orienting" system is assessed as the relative impairment observed when subjects have no idea where the upcoming target stimulus will appear (indexing their ability to quickly direct attention to that spatial location). The response-level "conflict monitoring" system is assessed as the impairment observed when aspects of the target stimulus is associated with incompatible responses, relative to only compatible responses.

The results

At the beginning of the experiment, the control and meditation training groups were matched for each subcomponent of attention, but the experienced meditators were superior to the other two groups in their ability to resist response conflict (thought to involve the dorsal system, perhaps the dorsal anterior cingulate, and perhaps the effects of concentrative meditation).

After training, only the retreat group showed a benefit in the ventral network, involved in monitoring of the environment, and consistent with the idea that development of this attentional subsystem requires more advanced meditation training. In particular, these subjects showed improvement in their ability to maintain vigilance in the absence of a warning about an upcoming target stimulus. Critically, the degree to which each subject showed this benefit was related to the number of months of previous experience they'd had with meditation.

Interestingly, the meditation-naive group which was trained in mindfulness meditation for 8 weeks showed an improvement in orienting relative to both the expert meditators and the control group. This again is consistent with the idea that this form of "concentrative attention" may be the first to improve with meditation training.

Jha et al conclude with some speculation as to the possible mechanisms for these effects. They suggest that during concentrative mindfulness meditation training, attention is repeatedly shifted back to the concentrated object of focus (usually breathing). This may exercise the same cognitive "shifting" process as is important in a variety of cognitive control tasks, such as task switching, but would do so in the context of a normal daily activity (breathing) and may therefore have more general and long-lasting effects by altering the function of the so-called "default network." Alternatively, meditation may improve cognition not by acting directly upon it, but by subduing an opponent process - such as via stress reduction.

[Sidenote: I learned Amishi Jha was my next-door neighbor in Philadelphia after getting into a streetside conversation with her husband about synthesizers. End sidenote.]