The Neuroscience of Intelligence: Empirical Support for the Theory of Multiple Intelligences?
C. Branton Shearer1 and Jessica M. Karanian2
1MI Research and Consulting
2 Department of Psychology, Boston College
Corresponding Author:
C. Branton Shearer
1316 S. Lincoln St.
Kent, OH 44240
Tel.: (330) 687-1735
E-mail:
C. Branton Shearer is the creator of the Multiple Intelligences Developmental Assessment Scales.
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Key Words: intelligence, multiple intelligences, cognition, general intelligence, neural correlates
Abstract
The concept of intelligence has been strongly debated since introduction of IQ tests in the 1900s. Numerous alternatives to unitary intelligence have achieved limited acceptance by both psychologists and educators. Multiple intelligences theory (Gardner, H. (1983,1993). Frames of mind: The theory of multiple intelligences. New York: Basic Books), despite criticism that it lacks empirical validity, has had sustained interest by educators worldwide. MI theory was one of the first to be based on neuroscience evidence. This investigation reviewed 318 neuroscience reports to conclude that there is robust evidence that each intelligence possesses neural coherence comparable with general intelligence. Implications for using MI theory as a bridge between instruction and cognitive neuroscience are discussed.
The concept of intelligence has a checkered history in the minds of many scientists and educational theorists. Many have abandoned the concept in part or entirely, and instead investigate cognitive abilities, problem-solving, or information processing capacities. However, many scientists have also investigated the functional neural systems that underlie intellectual achievement. The reason for this has been summed up succinctly by Jung and Haier [1, p. 171] “...there is no more important concept in education than the concept of intelligence”They assert that not all brains think the same way, thus “this simple fact could be revolutionary for education because it demands a neuroscience approach that recognizes the importance of individual differences and the necessity to evaluate each student as an individual” [2, p. 174].
The theory of multiple intelligences (MI) is of primary interest to the present investigation. Howard Gardner [3,4]redefined intelligence as the ability to solve problems or create products of value in a culture or community.Using this broad, common sense definition and eight criteria* that cover a range of evidence (e.g., neuroscience, workplace behaviors, great cultural achievements), Gardner identified eight distinct forms of intelligence that are possessed by all people, but in varying degrees. The eight intelligences identified are linguistic, logical-mathematical, spatial, kinesthetic, musical, interpersonal, intrapersonal and naturalist (for detailed descriptions, see Appendix A).
Traditional psychologists criticize MI theory for a number of reasons. One criticism is that MI theory lacks support from large scale studies [4,5] or experimental research[7,8,9. It has also been proposed that the eight intelligences are simply different manifestations of general intelligence [10,11]. An important practical criticism is that educators should not base instructional and curricular decisions upon a theory that lacks support from neuroscience evidence [12,13] and is unsubstantiated and unproven[14,15,16].
Amongneuroscientists, the predominant view on intelligence is that there iseither one general intelligence (g) or two types of intelligence (fluid and crystallized). However, there is a debate regardingthepossible sub-divisions of intelligence and each sub-division’s relationship to “g”. Numerous other theories that deviate from the unitary intelligence theory – includingtriarchic[17], emotional intelligence [18,19], structure of intellect [20], faculties of mind [21], and cognitive styles [22] – have had noteworthy, but limited, influence. Many have been recognized by the field of psychology, but not embraced by educators. Few have had the lasting and profound impact on education as multiple intelligences theory which is still of interest world-wide more than 30 years after its introduction [3, 4, 23]. Despite this broad appeal to educators, MI remains more of an inspirational educational framework rather than a fully developed scientific theory [24, 25, 2].
The practical critiques are of particular importance as the emerging field of educational cognitive neuroscience strives to establish a foundation for neuroscientific evidence-based instructional approaches. This new field has struggled to build practical connections between brain activity and instruction/curriculum. In its early years, there was widespread skepticism that brain-based education could develop without an explicit use of psycho-educational theory to bridge between neuronal activity and instruction [26]. This situation has improved more recently [27, 28, 29, 30], but the field continues to struggle to make a distinction between “pop psychology” of brain-based teaching and the science of educational cognitive neuroscience that can be systematically applied.
(Table 1 here)
The following literature review organizes 30 years of cognitive neuroscienceresearch on human cognition into core cognitive units that are each associated with a particular intelligence. We compared theneuroscientific evidence for each intelligence to the cortical areas outlined byGardner [3 ,4](Table 1) to address the following inter-related questions: (1) do these neural functional structures and networks display shared coherence while being conceptually unique and distinct from other functions, (2) taken together, do these data describe a solid conceptual framework for the “neural architecture” underlying each of the eight intelligences, and (3) how well do these neural architectures compare to what is known about the neural basis for general intelligence (i.e., g theory)? It should be underscored that this review of the cognitive neuroscience literature in relation to MI theory is intended to provide a foundation rather than a definitive examination of the constantly evolving literature on the neural underpinnings of human cognition.
Methods
Procedures
This investigation began with a detailed review of the various cognitive units and specific skills associated with each intelligence. For example, musical intelligence includes instrumental, vocal, composing and appreciation. Each of these ability sets includes technical skill as well as creative performance (e.g., singing on key and jazz improvisation) so the review of musical neuroscience studies would ideally be inclusive of this range of abilities. Charts were constructed for each intelligence with rows for MI Cognitive Units and columns for matched Neural Structures and Cognitive Skills (linguistic sample in Appendix B. All data is available upon request).
Using the terms related to each Cognitive Unit or specific skill (Table 2), PubMed or Google Scholar were used to search for published peer-reviewed empirical neuroscience studies (neural organization Appendix C and journals list in Appendix D). The goal was to identify a minimum of three to five studies per major skill area. Surprisingly, a great many more studies were obtained. Studies of personality characteristics or dispositions were not included (e.g., introversion, diligence, etc.). Theoretical articles or books were used mainly for background information. Several extensive meta-analysis and topic reviews served as guides to finding pertinent studies in the target area. Over 318 articles were referenced for the eight intelligences. The minimum number of studies was 19 for Logical-mathematical with a maximum of 73 for Intrapersonal (Table 2).
(Table 2 here)
From this wealth of knowledge excerpts from each text describing neural activations associated with carefully defined cognitive skills were entered into the charts per Cognitive Unit(see linguistic sample in Appendix B and E). As the investigation proceeded, the labels and defining characteristics for various Cognitive Units were adjusted to better align the neuroscience evidence with MI theory (Table 2, columns 6 and 7). This became a dialectical process between compatible perspectives. The next step was for an objective neuroscience doctoral student to review the data charts and harmonize the various neural descriptors according to standard neural anatomical terminology. All neural regions were then put into an Excel spreadsheet and reorganized based on neural hierarchy (Appendices C and E).
It became a challenge to manage the varieties of neural terminology. Neuroscientific researchers have used a wide variety of terms and labels and specificity over the years as the technology has evolved. Some researchers identified broad regions with a single label while others used multiple terms to identify sub-regions. Still others used Brodmann numbering,TalairachAtlas or the MNI Coordinate system. This variety of nomenclaturesrequired a careful translation and mapping onto the three-level hierarchy (Primary, sub-regions and particular structures) described below.
Our analysis of this data employed both qualitative and quantitative methods to determine if a three-dimensional view of the neural structures associated with each intelligence could be created. This hybrid approach – qualitative and quantitative – reflects both the evolution of the field as well as how the brain processes information – from very specific to diffuse patterns of activation.Studies were included in this analysis regardless of the type of the subjects employed to better reflect a wide variety of abilities. Some studies used undifferentiated subjects while others included those with brain damage and still others required the use of subjects with specifically defined skills.
Analyses
First, we assessed the frequency of cited primary neural regions, which included the frontal cortex, temporal cortex, parietal cortex, occipital cortex, cingulate cortex, insular cortex, subcortical regions, and the cerebellum. We also ran a secondary analysis on the primary regions that were most associated with each of the intelligences (i.e., primary regions that represented at least 20% of the primary neural citations). Within the top cited primary regions, we identified the top sub-regions. All sub-regions that represented at least 20% of a top primary neural regions were reported. Lastly, in some instances, a third-level analysis was conducted to identify the important sub-regions within a sub-region of a top primary neural region (e.g., frontal cortex prefrontal cortex dorsomedial prefrontal cortex; Appendix E). These second-level and third-level analyses are highlighted in the text.
Results
The following descriptions are highlights from an extensive dataset (see Appendix F). Complete data and interpretations are available as supplemental material.
Interpersonal
The interpersonal literature review identified 53 studies, including 111 citations of primary neural regions. The core cognitive units of interpersonal intelligence include social perception, interpersonal understanding, social effectiveness, and leadership. Results from the analysis of the primary neural regions can be found in Table 3 and Figure 1.
The analysis of primary neural regions revealed that interpersonal intelligence was most associated with the frontal cortex (43 citations). Secondary analyses more specifically identified that the prefrontal cortex (PFC) accounted for the large majority of frontal cortex citations (33/43 = 76.74%). A third-level analysis revealed that the dorsolateral PFC was the dominant sub-region within the PFC (8/33 = 24%).
Interpersonal intelligence was also associated with the temporal cortex as revealed by 31 citations. Within the temporal cortex, the medial temporal lobe (9/31 = 29%), amygdala (8/31 = 26%), and the superior temporal sulcus (7/31 = 23%) were the predominantly cited sub-regions. Other notable regions associated with Interpersonal intelligence included the cingulate cortex (12 citations), particularly the anterior cingulate cortex (ACC; 8/12 = 75%), and the parietal cortex (10 citations).
(Table 3 here)
(Figure 1 here)
Intrapersonal
The intrapersonal literature review identified 73 studies, including 219 citations of primary neural regions. The core cognitive units of intrapersonal intelligence include self-awareness, self-regulation, executive functions, and self-other management. Results from the analysis of the primary neural regions can be found in Table 4 and Figure 2.
The primary analysis revealed that Intrapersonal intelligence was most associated with the frontal cortex (90 citations) – the large majority of which were specific to the PFC (73/90 = 81%). A third-level analysis within the PFC revealed the dorsomedial PFC (18/73 = 25%) and the lateral PFC (15/73 = 21%) as major sub-regions.
The primary analysis also identified the cingulate cortex (37 citations), temporal cortex (36 citations), parietal cortex (25 citations), and subcortical regions (20 citations). Within the cingulate cortex, dominant sub-regions included the anterior cingulate cortex (27/37 = 73%). Within the temporal cortex, notable sub-regions included the medial temporal lobe (9/36 = 25%), amygdala (8/36 = 22%), and anterior temporal cortex (8/36 = 22%). Within the parietal cortex, the secondary analysis revealed that medial regions (10/25 = 40%) and inferior regions (5/25 = 20%) were dominant. Lastly, within the subcortical regions, the basal ganglia (10/20 = 50%) and brainstem (9/20 = 45%) were dominant. These structures are associated with cognition, learning, reward management, and unconscious memory (motor control).
(Table 4 here)
(Figure 2 here)
Visual-Spatial
The visual-spatial intelligence literature review identified 37 studies, including 143 citations of primary neural regions. The core cognitive units of visual-spatial intelligence include spatial cognition, working with objects, visual arts, and spatial navigation. Results from the analysis of the primary neural regions can be found in Table 5 and Figure 3.
The primary analysis revealed the frontal cortex as the most associated with visual-spatial intelligence (56 citations). Within the frontal cortex, secondary analyses identified the motor cortex (21/56 = 38%) and PFC (17/56 = 31%) as most important. A third-level analysis within the motor cortex highlighted the premotor cortex (12/21 = 57%) and the primary motor cortex (5/21 = 24%) as dominant. Within the PFC, the third-level analysis revealed the dorsolateral PC as most dominant (6/17 = 35%).
Furthermore, the primary analysis identified the parietal cortex (29 citations) as the second most dominant neural region for visual-spatial intelligence. Within the parietal cortex, the intraparietal sulcus (7/29 = 24%) and superior parietal lobule (7/29 = 24%) were notable sub-regions. A third-level analysis within the superior parietal lobule identified the precuneus as dominant (3/7 = 43%).
Other regions of interest included the temporal cortex (23 citations), including the medial temporal lobe (8/23 = 35%). A third-level analysis within the medial temporal lobe identified the hippocampus as the most dominant sub-region (4/8 = 50%). Furthermore, the primary analysis identified the occipital cortex (14 citations) as associated with visual-spatial intelligence, and a secondary analysis within the occipital cortex specifically identified the primary visual cortex as the most dominant sub-region (6/14 = 43%).
(Table 5 here)
(Figure 3 here)
Naturalist
The naturalist literature review identified 25 studies, including 58 citations of primary neural regions. The core cognitive units of naturalist intelligence derived from MI theory as well as the neuroscience literatureincluded pattern cognition, understanding living entities (including animals and plant life), and science. Typical behaviors that were studied include perceiving animal forms, motion, and vocalization; reading animal’s actions, intentions & emotions; biological life detection; and taxonomic thinking. No studies were found pertaining to understanding plant life. Results from the analysis of the primary neural regions can be found in Table 6 and Figure 4.
Analysis of the primary neural regions revealed that naturalist intelligence is most associated with the temporal cortex (19 citations). Within the temporal cortex, the secondary analysis identified the superior temporal sulcus (6/19 = 32%) and amygdala (5/19 = 26%) as notable.
The primary analysis also revealed subcortical neural regions (16 citations) as important for naturalist intelligence. Notable subcortical regions included regions of the brainstem (5/16 = 31%), the thalamus (5/16 = 31%), and the basal ganglia (4/16 = 25%).
(Table 6 here)
(Figure 4 here)
Musical
The musical literature review identified 42 studies, including 103 citations of primary neural regions. The core cognitive units of musical intelligence include music perception, music and emotions, and music production. Results from the analysis of the primary neural regions can be found in Table 7 and Figure 5.
Musical intelligence was most associated with the frontal cortex (42 citations). Within the frontal cortex, the motor cortex (31/42 = 74%) was the most dominant sub-region. A third-level analysis revealed the premotor cortex (12/31 = 39%) and the supplementary motor area (10/31 = 32%) as the most dominant sub-regions.
The next most frequently cited region was the temporal cortex (28 citations). A secondary analysis revealed the most notable sub-region was the superior temporal gyrus (23/28 = 82%), including the primary auditory cortex (19/23 = 83%, as revealed by a third-level analysis). Of other note, subcortical regions (16 citations) were also implicated, primarily accounted for by the basal ganglia (11/16 = 69%, as revealed by a third-level analysis).
(Table 7 here)
(Figure 5 here)
Kinesthetic
The kinesthetic literature review identified 41 studies, including 142 citations of primary neural regions. The core cognitive units of kinesthetic intelligence included body awareness and control, whole body movement, dexterity, and other types of movement (e.g., imitation, embodied cognition, gestures). Results from the analysis of the primary neural regions can be found in Table 8 and Figure 6.
The primary neural region analysis revealed the frontal cortex as most frequently cited (61 citations). A secondary analysis revealed that the dominant sub-region of the frontal cortex for kinesthetic intelligence was the motor cortex (46/61 = 75%). A third-level analysis further identified the primary motor cortex (19/46 = 41%), premotor cortex (15/46 = 33%), and supplementary motor area (9/46 = 20%) as dominant sub-regions.
Furthermore, the primary analysis identified the parietal cortex as the next most associated primary region (33 citations) within kinesthetic intelligence. Within the parietal cortex, the posterior parietal cortex was associated with the most citations (7/33 = 21%). Other regions of interest identified by the primary analysis included subcortical regions (15 citations), including the basal ganglia (11/15 = 73%, as indicated by secondary analysis) and thalamus (4/15 = 27%, as indicated by secondary analysis), as well as the cerebellum (13 citations).
(Table 8 here)
(Figure 6 here)
Linguistic
The linguistic literature review identified 28 studies, including 124 citations of primary neural regions. The core cognitive units of linguistic intelligence included speech, reading, writing, and communication. Results from the analysis of the primary neural regions can be found in Table 9 and Figure 7.