The actual title I gave this grant proposal for the Affective Neuroscience class was “Feature-specific, Antidepressant Mechanism of Transcranial Magnetic Stimulation”. But that is a bit of a mouthful for a blog post.

Introduction

Globally, 253 million people (3.6% of the world’s population) were affected by major depressive disorder (MDD) in 2013 (Vos et al., 2015). Other sources approximate annual prevalence rates at 5 to 15% (Lefaucheur et al., 2014). The American Psychiatric Association lists pharmacotherapy, psychotherapy and other somatic therapies (including electroconvulsive therapies) for acute treatment of MDD (Armstrong, 2011). Often, individuals with MDD are advised to undergo simultaneous psycho- and pharmacotherapy. Unfortunately, about half of cases of depression do not respond to antidepressants within 8 weeks of treatment onset, of which half may be accounted for by the placebo effect. Only an additional 8 % and 10% responded in the subsequent 8 weeks (DeRubeis et al., 2005). Remission rates within these 16 weeks peaked at 46%. Further, 10% or more of individuals with MDD are chronically unresponsive to pharmacological treatment (Berlim & Turecki, 2007).

Novel techniques must be developed to overcome the demonstrated unpredictability of traditional therapies. Early research implicates left prefrontal hypoactivation as a stable marker of depression; application of repetitive Transcranial Magnetic Stimulation (rTMS) to this area resulted in significant decrease in Hamilton depression rating scale (HDRS) scores (Pascual-Leone, Rubio, Pallardó, & Catalá, 1996). TMS is a non-invasive method for targeted neural activation using a coil that can induce magnetic fields. Repetitive TMS is a similar method with lower energy demands and that induces neural activation or inhibition depending on frequency of stimulation. Low-frequency rTMS (LF-rTMS) “inhibits” and high-frequency rTMS (HF-rTMS) “activates” cortical areas (Baeken & De Raedt, 2011).

Many studies have examined the relationship between mood disorders and asymmetries in the electroencephalographic (EEG) activity of the frontal cortex. This asymmetry may moderate or mediate emotional responding (Coan & Allen, 2004). In populations with MDD, left frontal activity is decreased and is associated with blunted reward processing. (Nusslock & Alloy, 2017). It is also suggested that the right frontal cortex is hyperactive(Lefaucheur et al., 2014). Together, frontal EEG asymmetry of alpha power between the left and right hemisphere is considered a biomarker for MDD (Nelson et al., 2014). Two main lines of treatments have been developed and researched: LF-rTMS to the right dorsolateral prefrontal cortex (DLPFC) [inhibit the hyperactivation] and HF- rTMS to the left DLPFC [excite the hypoactivation] (Lefaucheur et al., 2014).

Neural Correlates of Depression

The definition of depression began as a long-lasting imbalance of black bile and has become, in popular press, a “chemical imbalance” of neurotransmitters in the brain. This popular conception oversimplifies the disorder, which has a vast array of precursors and consequences. Concurrent with the improvement in brain imaging and observational technologies, psychopathology researchers are moving toward the use of neural profiles or bio-markers with the goal of personalizing treatment to an individual’s biosignature (Insel et al., 2010). In this framework, MDD has a set of behavioral features that can be correlated with stereotypical activation and activity patterns.

MDD presents with some or all of the following behaviors: anhedonia, anxiety (negative response bias), psychomotor retardation, disruption in sleep patterns and deficit in cognitive control (Drysdale et al., 2017).

Anhedonia, in unipolar depression, can be operationally defined as reduced approach motivation along with reduced positive affect (Nusslock & Alloy, 2017). In a behavioral sense, the motivation to work for rewards and learning reward cues is negatively affected. This feature of MDD is associated with the fronto-striatal reward circuit. This circuit contains dopaminergic projections from the ventral tegmental area (VTA) to the nucleus accumbens (NAcc) of the ventral striatum, which processes reward-related properties of a stimulus, to cortical regions (Hamilton et al., 2012; Dillon et al., 2015). It has been shown in animals that there are descending pathways from prefrontal cortex (PFC) that modulate dopaminergic activity of subcortical areas such as the striatum (Strafella, Paus, Barrett, & Dagher, 2001). Individuals with MDD present with less ventral striatal response than healthy comparisons during reward anticipation and outcome (Forbes et al., 2009; Dillon et al., 2015; Nusslock & Alloy, 2017). Further, altered dopaminergic functioning in the NAcc robustly prevents the organism’s pursuit of the rewarding stimulus (Salamone, Correa, Farrar, & Mingote, 2007) Depressed individuals show weaker reward learning signals in the ventral striatum, dorsal anterior cingulate cortex (ACC) and dorsal caudate, which instantiate reinforcement learning of reward cues (Dillon et al., 2015). Though these findings are preliminary, they suggest that anhedonia is associated with dysfunction in the fronto-striatal reward circuit in developing response biases that present along with reduced pursuit of reward (Dillon et al., 2015). Recent work on EEG alpha power asymmetry in individuals with MDD has found that the driving force in that relationship is the underlying approach motivation deficits (Nelson et al., 2014).

Though more stereotypical of the Generalized Anxiety Disorder, anxiety and threat vigilance manifests in some individuals with MDD as slower disengagement with negative stimuli and cognition (Dillon et al., 2015). A meta-analysis of functional neuroimaging studies revealed that populations with MDD show greater activation in the amygdala, insula and dorsal ACC and lower activation in dorsal striatum and DLPFC in response to negative stimuli compared to healthy subjects (Hamilton et al., 2012). It also found higher baseline activity in bilateral pulvinar nuclei of the thalamus. These results suggest an increased salience of negative information in MDD. Due to concomitant reduced striatal dopamine levels, there are diminished responses to negative stimuli in dorsolateral PFC, thus not initiating the attributions and appraisals that are typical of healthy individuals (Hamilton et al., 2012).

Another feature of MDD that has implicated neural underpinnings is psychomotor retardation, typified by slowing of motility, mental activity and speech (Bennabi, Vandel, Papaxanthis, Pozzo, & Haffen, 2013). Patients with MDD and psychomotor retardation are shown to have reduced extracellular dopamine in the caudate and putamen. The basal ganglia is another candidate for irregularity in MDD as it has been shown to uniquely predict psychomotor speed (Naismith et al., 2002).

Dysfunction in cognitive control is a feature of MDD that operates on the previously mentioned feature. Through projections to mesolimbic and limbic structures, PFC activity has a modulatory influence on lower structures (Wood & Grafman, 2003). Reduced activity in the left dosolateral and dorsomedial PFC has been documented in depressed patients (Davidson, Lewis, et al., 2002). This is, in part, due to reduced gray matter volume in this region as shown in MRI-derived morphometric measures (Drevets et al., 1997). The dorsolateral PFC has been implicated in attention allocation, interference resolution and information represention (Davidson, Pizzagalli, Nitschke, & Putnam, 2002; Wood & Grafman, 2003; Hamilton et al., 2012), acting as a part of an executive/control network (Liston et al., 2014). As mentioned previously, disinhibition from PFC in MDD obstructs information representation from contextualization and reappraisal (Hamilton et al., 2012). This presents as increased salience of negative stimuli in depressed individuals.

The ventromedial PFC works through a similar mechanism with the amygdala. Relative to neurologically healthy subjects, patients with ventromedial PFC lesions exhibited potentiated amygdala responses to aversive images (Motzkin, Philippi, Wolf, Baskaya, & Koenigs, 2015). This mechanism would play a greater role in the anxiety components of MDD.

A recent study of resting-state connectivity of multi-site samples showed that patients with MDD can be clustered based on similar patterns of dysfunctional connectivity in frontostriatal and limbic networks that are associated with different behavior features of the disorder (Drysdale et al., 2017). These researchers found common pathalogical profiles amongst the insula, orbitofrontal cortex, VMPFC and other subcortical areas. Superimposed on the shared pathologies, different clusters of individuals had distinct patterns of abnormalities in anhedonic- and anxiety-, insomniac- and psychomotor-related functional connectivity. Hyperconnectivity in the thalamic and frontostriatal networks for reward processing and action initiation factored into the anhedonic-related connectivity score. Hypoconnectivity in the frontoamygdala networks for fear-related behaviors factored into the anxiety-related connectivity score. Reduced connectivity in ACC and orbitofrontal areas for motivation and incentive-salience evaluation was characterized by anergia and fatigue (Drysdale et al., 2017). Further, these patterns of connectivity predicted responsiveness of an individual to rTMS. For instance, the biotype typified by increased anxiety and anergia/fatigue was most responsive (82%) to HF-rTMS to dorsomedial PFC (DMPFC). These results are a starting point for further investigation into refined parameterization of rTMS treatment based on connectivity profiles that map to profiles of the behavioral features of depression.

TMS Technology

Barker et al. (Barker, Jalinous, & Freeston, 1985) was the first application of Faraday’s basic law of electromagnetism in order to stimulate regions of the brain. Subsequently, the clinical method of TMS was established and has become a routine procedure for use in various neurological and psychiatric disorders (Lefaucheur et al., 2014).

A high current pulse generator (up to several thousand amperes) discharges current through a stimulating coil, producing a brief magnetic pulse with strengths up to several Teslas. When the coil is placed onto a scalp, the generated magnetic field induces an electric field sufficient to polarize or depolarize axons in order to activate or inhibit neural networks in the cortex. There are many dependencies of the coil’s action on physical and biological parameters such as type of coil, orientation of coil, head geometry, magnetic pulse waveform, the intensity, frequency and pattern of stimulation and relative orientation of neural elements to magnetic fields produced (Lefaucheur et al., 2014). TMS can activate or inhibit networks depending on the setting of the above parameters. LF-rTMS (Æ 1 Hz) “inhibits” cortical activity, while high-frequency rTMS (HF-rTMS) (> 1 Hz) “activates” cortical areas (Baeken & De Raedt, 2011; Chervyakov, Chernyavsky, Sinitsyn, & Piradov, 2015).

TMS as an Anti-Depressant

The ability for rTMS to influence mood was accidentally discovered and was demonstrated as an alternative to convulsive therapies in cases of drug-resistant depression (Pascual-Leone et al., 1996). For depression, the typical treatment is HF-rTMS with a biphasic waveform to the left DLPFC. This technique is used in cases of depression that do not respond to pharmocological agents, but only some patients show a positive outcome (Baeken & De Raedt, 2011; Lefaucheur et al., 2014). Nevertheless, the American Psychiatric Association, the Canadian Network for Mood and Anxiety Treatments, and the World Federation of Societies of Biological Psychiatry endorse rTMS as a treatment for MDD. Though rTMS has been shown to be relatively effective in treating medication-resistant or refractory depression, the underlying, neurophysiological mechanisms of these effects are largely unknown (Baeken & De Raedt, 2011; Dayan, Censor, Buch, Sandrini, & Cohen, 2013; Chervyakov et al., 2015). Most treatments and studies of treatment involve HF-rTMS to the left DLPFC or LF-rTMS to the right DLPFC due to distance of effect of the method. Other methods (such as deep brain stimulation) must be used to reach deeper targets such as the VMPFC or mesolimbic structures (Dayan et al., 2013). One common finding is that rTMS reduces connectivity and heightens metabolic activity in the ACC in individuals with MDD (Teneback et al., 1999; Paus, Castro-alamancos, & Petrides, 2001; Li et al., 2010; Baeken & De Raedt, 2011; Liston et al., 2014), bringing its action closer to that of a healthy control. Thus, baseline ACC activity is often a predictor of rTMS efficacy.

At lower levels of analysis, rTMS also has an effect through prevention of neuronal death, dendritic growth and concentration of brain-derived neurotrophic factors (Dayan et al., 2013; Chervyakov et al., 2015). These local changes affect long-distance functional connectivities.

The induced effects of rTMS outlast the period of stimulation. Repetitive electrical stimulation in animal and human models have induced long-term potentiation (LTP) and depression (LTD) -like plasticity (Dudek & Bear, 1993; Huang, Edwards, Rounis, Bhatia, & Rothwell, 2005). A pre- and post-treatment connectivity analysis revealed that HF-rTMS to the left DLPFC modulated functional connectivity of cortical networks, following Hebb’s Law. Specifically, hyperconnectivity of the subgenual ACC (sgACC) with the default mode network (implicated in rumination and self-referential processing) was reduced. Yet, the abnormalities of the executive or control network (implicated in directing attention and decision making) persisted even though the target of the stimulation (DLPFC) is contained within the central executive network. They proposed that the former effect was produced through demonstrated interactions between the two networks (Liston et al., 2014). Connectivity profiles of the sgACC was shown to predict responsiveness of an individual to left DLPFC HF-rTMS treatment for MDD in this (Liston et al., 2014) and another study (Li et al., 2010). Further, more elaborate connectivity profiles were also shown to predict responsiveness to DMPFC HF-rTMS treatment (Drysdale et al., 2017).

Proposal

Drysdale et. al. 2017 (Drysdale et al., 2017) found preliminary evidence that connectivity profiles in the frontostriatal, frontoamydala and thalamic networks as well as in the ACC and orbitofrontal cortex may map to biotypes of depression that are characterized by graded ratings of different features of MDD. Research concerning the structure-level networks in individuals with MDD have implicated a dependency of rTMS treatment efficacy on these connectivity profiles that may instantiate the specific behavioral phenotype of their disorder. Here, I propose an integrative, computational study that asks whether connectivity quantities can be used to predict the downstream effect of rTMS for individuals with MDD.

Resting state functional magnetic imaging will be used to measure functional connectivity pre- and post-treatment. Structures implicated in feature-related behaviors previously outlined will be seeded for analysis. Treatment will be HF-rTMS applied to the left DLPFC for all subjects. The study will require 50 participants with MDD and 50 matched controls.

Structures of the frontostriatal circuit, including the left DLPFC, NAcc of the ventral striatum, ventral tegmental area, dorsal ACC and caudate, will be seeded and assumed as factors in anhedonia-related behaviors. Structures of the frontoamygdala circuit, including but not limited to the VMPFC, amygdala, insula, dorsal ACC, and dorsal striatum, will be seeded and assumed as factors in anxiety-related behaviors. Incentive-salience and motivation networks will also be seeded and associated with the psychomotor and anergia-related behaviors. Structures to be seeded in these networks will be chosen based on further research.

Additionally, each individual will have the features of their MDD evaluated using the Hamilton Rating Scale for Depression before and after treatment. This scale distinguishes the features of MDD that are potentially correlated with the connectivities outlined above, including anhedonia, anxiety and psychomotor- or anergia-related behaviors. This scale does not isolate salience-related features, which must be behaviorally measured with another method.

After all pre- and post-treatment imaging is complete, an analysis of rTMS efficacy will be conducted. Using computational and machine learning methods, we will examine whether, given pre-treatment connectivities and rTMS treatment location/intensity, post-treatment connectivities may be predicted. The model will have prior knowledge about influence of one neural structure to another in order to guide parameterization. Further refinement of this technique may lead to the development of a clinical test that determines whether rTMS would have an effect on the connectivity profile of an individual with MDD. Relationships between behavioral changes and connectivity changes will be examined in order to understand how the designated networks play a role in the presentation of their associated behaviors.

Together, these results will provide some evidence for the mechanism of rTMS on connectivity profiles of individuals with MDD. It also has the potential to confirm the associations of particular depressive feature categories with structural networks that have been previously implicated. Lastly, experimenting with using connectivity profiles to predict rTMS effectiveness may lead to a clinical test that guides treatment application.

Citations

Armstrong, C. (2011). APA Releases Guidelines on Treatment of Patients with Major Depressive Disorder. Am. Psychiatr. Assoc., 83(10), 1219–1227.

Baeken, C., & De Raedt, R. (2011). Neurobiological mechanisms of repetitive transcranial magnetic stimulation on the underlying neurocircuitry in unipolar depression. Dialogues Clin Neurosci, 13(1), 139–145.

Barker, A., Jalinous, R., & Freeston, I. (1985). NON-INVASIVE MAGNETIC STIMULATION OF HUMAN MOTOR CORTEX. The Lancet, 325(8437), 1106 - 1107. doi: http://dx.doi.org/10.1016/S0140-6736(85)92413-4

Bennabi, D., Vandel, P., Papaxanthis, C., Pozzo, T., & Haffen, E. (2013). Review Article Psychomotor Retardation in Depression : A Systematic Review of Diagnostic , Pathophysiologic , and Therapeutic Implications. Biomed Res., 2013.

Berlim, M. T., & Turecki, G. (2007). Treatment-Resistant Refractory Major Depression : A Review of Current Concepts and Methods. La Rev. Can. Psychiatr., 52(1).

Chervyakov, A. V., Chernyavsky, A. Y., Sinitsyn, D. O., & Piradov, M. A. (2015). Possible Mechanisms Underlying the Therapeutic Effects of Transcranial Magnetic Stimulation. Front. Hum. Neurosci., 9(June), 303. doi: 10.3389/fnhum.2015.00303

Coan, J. A., & Allen, J. J. B. (2004). Frontal EEG asymmetry as a moderator and mediator of emotion. Biol. Psychol., 67(1-2), 7–49. doi: 10.1016/j.biopsycho.2004.03.002

Davidson, R. J., Lewis, D. A., Alloy, L. B., Amaral, D. G., Bush, G., Cohen, J. D., . . . Peterson, B. S. (2002). Neural and behavioral substrates of mood and mood regulation. Biol. Psychiatry, 52(6), 478–502. doi: 10.1016/S0006-3223(02)01458-0

Davidson, R. J., Pizzagalli, D., Nitschke, J. B., & Putnam, K. (2002). Depression: perspectives from affective neuroscience. Annu. Rev. Psychol., 53, 545–574. doi: 10.1146/annurev.psych.53.100901.135148

Dayan, E., Censor, N., Buch, E. R., Sandrini, M., & Cohen, L. G. (2013). Noninvasive brain stimulation: from physiology to network dynamics and back. Nat. Neurosci., 16(May 2016), 838–44. doi: 10.1038/nn.3422

DeRubeis, R. J., Hollon, S. D., Amsterdam, J., Shelton, R. C., Young, P. R., Salomon, R. N., . . .

Gallop, R. (2005). Cognitive Therapy vs Medications in the Treatment of Moderate to Severe Depression. Arch. Gen. Psychiatry, 62, 409–416. doi: http://dx.doi.org/10.1001/archpsyc.62.4.409

Dillon, D. G., Rosso, I. M., Pechtel, P., Killgore, W. D. S., Rauch, S. L., & Pizzagalli, D. A. (2015). Responses and Reward Processing in Anxiety and Depression. Depress Anxiety, 31(3), 233–249. doi: 10.1002/da.22202.Peril

Drevets, W. C., Price, J. L., Simpson Jr., J. R., Todd, R. D., Reich, T., Vannier, M., & Raichle, M. E. (1997). Subgenual prefrontal cortex abnormalities in mood disorders. , 386(6627), 824–827. doi: 10.1038/386824a0

Drysdale, A. T., Grosenick, L., Downar, J., Dunlop, K., Mansouri, F., Meng, Y., . . . Liston, C. (2017). Resting-state connectivity biomarkers define neurophysiological subtypes of depression. Nat. Publ. Gr., 23(1). doi: 10.1038/nm.4246

Dudek, S. M., & Bear, M. F. (1993). Bidirectional long-term modification of synaptic effectiveness in the adult and immature hippocampus. J. Neurosci., 13(7), 2910–8. doi: 10.1038/305719a0

Forbes, E. E., Hariri, A. R., Martin, S. L., Silk, J. S., Moyles, D. L., Fisher, P. M., . . . Dahl, R. E. (2009). Altered striatal activation predicting real-world positive affect in adolescent major depressive disorder. Am. J. Psychiatry, 166(1), 64–73. doi: 10.1176/appi.ajp.2008.07081336

Hamilton, J. P., Etkin, A., Furman, D. J., Lemus, M. G., Johnson, R. F., & Gotlib, I. H. (2012). Functional neuroimaging of major depressive disorder: a meta-analysis and new integration of base line activation and neural response data. Am. J. Psychiatry, 169(7), 693–703. doi: 10.1176/appi.ajp.2012.11071105

Huang, Y. Z., Edwards, M. J., Rounis, E., Bhatia, K. P., & Rothwell, J. C. (2005). Theta burst stimulation of the human motor cortex. Neuron, 45(2), 201–206. doi: 10.1016/j.neuron.2004.12.033

Insel, T., Cuthbert, B., Garvey, M., Heinssen, R., Pine, D. S., Quinn, K., . . . Wang, P. (2010). Research Domain Criteria (RDoC): Toward a new classification framework for research on mental disorders. Am. J. Psychiatry, 167(7), 748–751. doi: 10.1176/appi.ajp.2010.09091379

Lefaucheur, J.-P. P., Andre-Obadia, N., Antal, A., Ayache, S. S., Baeken, C., Benninger, D. H., . . . Garcia-Larrea, L. (2014). Evidence-based guidelines on the therapeutic use of repetitive transcranial magnetic stimulation (rTMS). Clin Neurophysiol, 125(11), 2150–2206. doi: S1388-2457(14)00296-X [pii]1 ̊0.1016/j.clinph.2014.05.021 [doi]

Li, C. T., Wang, S. J., Hirvonen, J., Hsieh, J. C., Bai, Y. M., Hong, C. J., . . . Su, T. P. (2010). Antidepressant mechanism of add-on repetitive transcranial magnetic stimulation in medication-resistant depression using cerebral glucose metabolism. J. Affect. Disord., 127(1-3), 219–229. doi: 10.1016/j.jad.2010.05.028

Liston, C., Chen, A. C., Zebley, B. D., Drysdale, A. T., Gordon, R., Leuchter, B., . . . Dubin, M. J. (2014). Default mode network mechanisms of transcranial magnetic stimulation in depression. Biol. Psychiatry, 76(7), 517–526. doi: 10.1016/j.biopsych.2014.01.023

Motzkin, J. C., Philippi, C. L., Wolf, R. C., Baskaya, M. K., & Koenigs, M. (2015). Ventromedial prefrontal cortex is critical for the regulation of amygdala activity in humans. Biol. Psychiatry, 77(3), 276–284. doi: 10.1016/j.biopsych.2014.02.014

Naismith, S., Hickie, I., Ward, P. B., Turner, K., Scott, E., Little, C., . . . Parker, G. (2002). Caudate nucleus volumes and genetic determinants of homocysteine metabolism in the prediction of psychomotor speed in older persons with depression. Am. J. Psychiatry, 159(12), 2096–2098. doi: 10.1176/appi.ajp.159.12.2096

Nelson, B. D., Mcgowan, S. K., Sarapas, C., Robison-andrew, E. J., Altman, E., Campbell, M. L., . . . Stewart, A. (2014). NIH Public Access. J Abnorm Psychol, 122(3), 662–671. doi: 10.1037/a0033982.Biomarkers

Nusslock, R., & Alloy, L. B. (2017). Reward Processing and Mood-Related Symptoms: An RDoC and Translational Neuroscience Perspective. J. Affect. Disord.(June 2016), 0–1. doi: 10.1016/j.jad.2017.02.001

Pascual-Leone, A., Rubio, B., Pallardó, F., & Catalá, M. D. (1996). Rapid-rate transcranial magnetic stimulation of left dorsolateral prefrontal cortex in drug-resistant depression. Lancet, 348(9022), 233–237. doi: 10.1016/S0140-6736(96)01219-6

Paus, Ï., Castro-alamancos, M. a., & Petrides, M. (2001). SHORT COMMUNICATION Cortico-cortical connectivity of the human mid-dorsolateral frontal cortex and its modulation by repetitive transcranial magnetic stimulation. Eur. J. Neurosci., 14, 1405–1411.

Salamone, J. D., Correa, M., Farrar, A., & Mingote, S. M. (2007). Effort-related functions of nucleus accumbens dopamine and associated forebrain circuits. Psychopharmacology (Berl)., 191(3), 461–482. doi: 10.1007/s00213-006-0668-9

Strafella, A. P., Paus, T., Barrett, J., & Dagher, A. (2001). Repetitive transcranial magnetic stimulation of the human prefrontal cortex induces dopamine release in the caudate nucleus. J. Neurosci., 21(September), 1–4. doi: 20015457 [pii]

Teneback, C. C., Nahas, Z. H., Speer, A. M., Molloy, M., Stallings, L. E., Spicer, K. M., . . . George, M. S. (1999). Changes in prefrontal cortex and paralimbic activity in depression following two weeks of daily left prefrontal TMS. J. Neuropsychiatry Clin. Neurosci., 11(4), 426–435. doi: 10.1176/jnp.11.4.426

Vos, T., Barber, R. M., Bell, B., Bertozzi-Villa, A., Biryukov, S., Bolliger, I., . . . Murray, C. J. L. (2015). Global, regional, and national incidence, prevalence, and years lived with disability for 301 acute and chronic diseases and injuries in 188 countries, 1990-2013: A systematic analysis for the Global Burden of Disease Study 2013. Lancet, 386(9995), 743–800. doi: 10.1016/S0140-6736(15)60692-4

Wood, J. N., & Grafman, J. (2003). Human prefrontal cortex: processing and representational perspectives. Nat. Rev. Neurosci., 4(2), 139–147. doi: 10.1038/nrn1033